Vehicles for Semantic Inheritance - University of … · The initial interpretational steps include protein synthe- ... velopments of abstraction in writing, algorithms, ... tional

Life, Information, Entropy,and TimeVehicles for Semantic Inheritance

Attempts to understand how information content can be included in an accounting ofthe energy flux of the biosphere have led to the conclusion that, in informationtransmission, one component, the semantic content, or “the meaning of the message,”adds no thermodynamic burden over and above costs arising from coding, transmis-sion and translation. In biology, semantic content has two major roles. For all lifeforms, the message of the genotype encoded in DNA specifies the phenotype, and hencethe organism that is tested against the real world through the mechanisms of Dar-winian evolution. For human beings, communication through language and similarabstractions provides an additional supra-phenotypic vehicle for semantic inheri-tance, which supports the cultural heritages around which civilizations revolve. Thefollowing three postulates provide the basis for discussion of a number of themes thatdemonstrate some important consequences. (i) Information transmission througheither pathway has thermodynamic components associated with data storage andtransmission. (ii) The semantic content adds no additional thermodynamic cost. (iii)For all semantic exchange, meaning is accessible only through translation and inter-pretation, and has a value only in context. (1) For both pathways of semanticinheritance, translational and copying machineries are imperfect. As a consequenceboth pathways are subject to mutation and to evolutionary pressure by selection.Recognition of semantic content as a common component allows an understanding ofthe relationship between genes and memes, and a reformulation of Universal Dar-winism. (2) The emergent properties of life are dependent on a processing of semanticcontent. The translational steps allow amplification in complexity through combina-torial possibilities in space and time. Amplification depends on the increased potentialfor complexity opened by 3D interaction specificity of proteins, and on the selection ofuseful variants by evolution. The initial interpretational steps include protein synthe-sis, molecular recognition, and catalytic potential that facilitate structural and func-tional roles. Combinatorial possibilities are extended through interactions of increas-ing complexity in the temporal dimension. (3) All living things show a behavior thatindicates awareness of time, or chronognosis. The �4 billion years of biologicalevolution have given rise to forms with increasing sophistication in sensory adapta-tion. This has been linked to the development of an increasing chronognostic range,and an associated increase in combinatorial complexity. (4) Development of a modernhuman phenotype and the ability to communicate through language, led to thedevelopment of archival storage, and invention of the basic skills, institutions andmechanisms that allowed the evolution of modern civilizations. Combinatorial am-plification at the supra-phenotypical level arose from the invention of syntax, gram-

Antony R. Crofts

Antony R. Crofts is at the Departmentof Biochemistry, 419 Roger AdamsLaboratory, 600 S. Mathews Avenue,University of Illinois at Urbana-Champaign, Urbana, IL 61801; E-mail:[email protected]

14 C O M P L E X I T Y © 2007 Wiley Periodicals, Inc., Vol. 13, No. 1DOI 10.1002/cplx.20180

Published online 8 October 2007 in Wiley InterScience(www.interscience.wiley.com)

mar, numbers, and the subsequent de-velopments of abstraction in writing,algorithms, etc. The translational ma-chineries of the human mind, the “mu-tation” of ideas therein, and the “conver-sations” of our social intercourse, haveallowed a limited set of symbolic de-scriptors to evolve into an exponentiallyexpanding semantic heritage. (5) Thethree postulates above open interestingepistemological questions. An under-standing of topics such dualism, the elanvital, the status of hypothesis in science,memetics, the nature of consciousness,the role of semantic processing in thesurvival of societies, and Popper’s threeworlds, require recognition of an insub-stantial component. By recognizing anecessary linkage between semantic con-tent and a physical machinery, we canbring these perennial problems into theframework of a realistic philosophy. It issuggested, following Popper, that the �4billion years of evolution of the bio-sphere represents an exploration of thenature of reality at the physicochemicallevel, which, together with the consciousextension of this exploration throughscience and culture, provides a firm epis-temological underpinning for such aphilosophy. © 2007 Wiley Periodicals,Inc. Complexity 13: 14 –50, 2007

Key Words: DNA; genotype; evolution;chronognosis; semantic content; Pop-per; evolutionary epistemology; memes

INTRODUCTION

T he arguments of this essay arisefrom two apparently unrelatedthemes. The first of these is an at-

tempt to address the question of how tointegrate the information content of thebiosphere into quantification of globalenergy flux. The second is the evolutionof temporal perception and its extensioninto consciousness. Somewhat surpris-ingly, these areas coalesce because of apreviously unrecognized property of onecomponent of information content. AsShannon recognized, communication re-quires two components—a thermody-namic framework for coding and trans-

mission, and the semantic content—the“meaning in the message.” I will arguethat this latter component adds no addi-tional thermodynamic burden. This ther-modynamic inconsequence raises ques-tions about how we know about theworld, echoing the epistemological con-cerns of philosophers down the ages. Myfirst task therefore is to explain why thereis a problem in incorporating informationcontent into the framework throughwhich we perform our energy account-ing. The second line of enquiry intro-duces recognition of the importance oftemporal perception in defining behaviorand the linkage between sophistication insensory perception, and awareness oftime—a theme that leads into the peren-nial philosophical enquiry about the na-ture of the mind. The two come togetherin our understanding of the role of infor-mation transmission in evolution and incivilization, of the emergence of life andof consciousness, and of how our presentview of the world has evolved over thepast few millennia. What comes out ofthis enquiry is a fresh perspective on top-ics such as the emergence of biologicalcomplexity, the duality of mind and body,consciousness, and the “vital force.” Foreach of these, some aspects have beenrejected from consideration by conven-tional science, but they continue to inter-est philosophers. They can be broughtinto a scientific context by recognition ofa necessary coupling between the twocomponents of communication recog-nized by Shannon: the “engineeringproblem” and semantic content.

INTERCEPTION OF SOLAR ENERGY BYTHE BIOSPHEREEarth intercepts a small fraction of solarradiation, determined by simple geomet-

ric parameters. Since, in the aggregate,earth remains in steady state, the energyloss must balance the energy input. Theuniverse pays its second-law dues by re-leasing a larger number of quanta, mainlyin the IR, each at a lower energy appro-priate to the differences in temperaturebetween sun and earth. The differentfluxes contributing to this overall processare quite complex. Major fractions in-clude reflection and scattering by clouds,polar ice, bare land, and water; absorp-tion by the atmosphere with energy lossthrough thermal (infrared emission) andchemical pathways; passive absorptionand re-emission of IR by bare land; ab-sorption by water, with direct re-emissionas IR, or after a delay through the hydro-logic cycle; and absorption by plants. Theoverall energy balance also includes thethermal emission from earth’s endoge-nous heat sources.

Where does life come into the equa-tion? As recognized by Boltzmann asearly as 1886

The general struggle for existenceof living beings is therefore not afight . . . for energy, which is plen-tiful in the form of heat, unfortu-nately untransformably, in everybody. Rather, it is a struggle forentropy . . . that becomes avail-able through the flow of energyfrom the hot Sun to the coldEarth. To make the fullest use ofthis energy, the plants spread outthe immeasurable areas of theirleaves and harness the Sun’s en-ergy by a process as yet unex-plored, before it sinks down tothe temperature level of ourEarth, to drive chemical synthe-ses of which one has no inkling asyet in our laboratories. Broda [1]

Although the mechanism of photosyn-thesis is now well explored, and we havemore than an “inkling” of the chemicalprocesses, Boltzmann’s insight and gen-eral heat-engine treatment are still ap-propriate [2, 3]. If we assume that the

What comes out of this enquiry isa fresh perspective on topics

such as the emergence ofbiological complexity, the dualityof mind and body, consciousness,

and the “vital force.”

© 2007 Wiley Periodicals, Inc. C O M P L E X I T Y 15DOI 10.1002/cplx

biosphere is, over an appropriate time in-terval, in steady state, any energy inter-cepted by the thin layer of life at the surfaceof earth must also eventually be convertedto heat at terrestrial temperatures. As a con-sequence, since the cross-sectional area forsolar interception is the same, the overallenergy equation would be unaltered. Fac-tors influencing albedo and emission, suchas changes in the greenhouse effect, willcomplicate this simple picture, but we willignore these for the moment. A substantialfraction of the incoming light energy is in-tercepted by the biosphere, predominantlyby green plants, algae, and cyanobacteria.The intercepted fraction provides the inputfor photosynthesis, and hence the maindriving force that sustains life on earth. Es-timations of the fraction diverted to the bio-sphere vary somewhat, but the fraction issignificant. Of the solar flux intercepted bythe planet (the solar constant of �1379 Wm�2 of projected area) about 50% reachesthe surface (Figure 1). On land about 33%arrives during a growing season, and about20% of that is intercepted by leaves. Addi-tional losses due to reflection (20%) and apoor color match (50%) mean that onlyabout 0.8% of the light intercepted is savedas photosynthetic product. Of that, about40% is used for maintenance, so �0.5% ofthe solar flux at the land surface is availableto fuel the other side of the biosphere: an-imal consumption, bacterial and fungal re-cycling. The productivity in the oceans isbuffered seasonally, so that, although theinstantaneous flux doesn’t match that ofthe land, the gross annual ocean productiv-ity is about the same as that of the land. It isestimated that about 100 billion metric tonsof carbon dioxide are assimilated annuallythrough photosynthesis, corresponding to3 � 1018 kJ (compared with 4.35 � 1017 kJworld energy consumption by man in2002) [4, 5]. This ratio is usually used toillustrate the scale of the photosynthetic en-ergy system, but it is salutory to realize thatalready the anthropogenic contribution is 1

7that of the rest of the biosphere.

HARVESTING THE ENTROPIC YIELD:THE OTHER GREAT REVOLUTION INBIOLOGYOur knowledge of the energy fluxesthrough the biosphere is still far fromcomplete. However, the broad outlines

are well understood, suggesting that in

principle we could quantify terms nec-

essary for a more complete thermody-

namic description. Much recent atten-

tion has been given to this area in the

context of attempts to predict the ef-

fects of increased CO2 and global warm-

ing on biomass yield, crop productivity,

etc. [4, 5], but major advances in our

understanding of the molecular mech-

anisms involved provide the bedrock on

which these efforts are based.

When photons are absorbed, they

excite transitions in absorber bands

matched in energy, which decay

through photochemistry, or through

emission of fluorescence, phosphores-

cence, and/or of IR photons. For most

materials, the IR emission dominates.

Some of the photons are emitted more

or less instantaneously; others are re-

tained by absorption in lower vibronic

bands, leading to a temperature change

and exchange of heat with the environ-

ment, and are re-emitted in the IR after

a delay, the so-called latent heat. For

absorption of radiation incident on the

earth, the timescales for these processes

range from “instantaneous” scattering

with minimal loss of energy, to picosec-

onds for internal conversion, up to the

diurnal and annual (or longer) cycles

involving latent heat, the latter reflect-

ing, for example, redistribution of ther-

mal energy through winds and current,

and the cyclic loss and accumulation of

ice at the poles. Without biological in-

tervention, these decay processes are

uncoupled from any complementary

process through which the potential

work content can be conserved; they

are completely dissipative. Some ab-

sorption by water leads to evaporation

and a different sort of delay through the

hydrologic cycle, with a range of time

scales that can be much longer. This

delay is due in part to a coupling of the

energy absorbed to evaporative work

against gravity, and against the osmotic

potential of salt waters. However, this

conserved energy is eventually also dis-

sipated without further coupling.

In contrast to the dissipative processes

above, absorption of photons by photo-

synthetic systems introduces pathways

for energy conversion that are directly

FIGURE 1

Scheme to show how the solar energy intercepted by the earth is redistributed. The photosyntheticyield of �0.5% is available for consumption by animals, fungi, and bacteria, and sustains thebiosphere, but is eventually lost to space as IR radiation (Adapted from http://asd-www.larc.nasa.gov/erbe/components2.gif). [Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]

16 C O M P L E X I T Y © 2007 Wiley Periodicals, Inc.DOI 10.1002/cplx

coupled to photochemical, chemical andelectrochemical reactions through whichsome of the work content is conserved [2,3]. These initiate the chain of reactionsthrough which the biosphere is sustained.Our understanding of these reactions hasdeveloped remarkably over the last fewdecades, but is only part of a larger revo-lution in biology. One of the triumphs ofthe biological sciences over the last 70years has been the elucidation of the met-abolic web of coupled reactions throughwhich the energy content of light or foodis conserved and used to drive the syn-thetic processes through which life ismaintained. After the introduction ofMitchell’s chemiosmotic hypothesis[6–10] in 1961, elucidation of the mecha-nisms of the main energy transductionprocesses of the biosphere, photosynthe-sis and respiration, has revealed a wholeset of electrochemical mechanisms cou-pling photochemical and electron trans-fer reactions to transport of protons (or ina few cases Na�) across membranes.From a combination of biophysical, bio-chemical, structural, and molecular engi-neering studies, these reactions are inmany cases now understood at the mo-lecular level [11]. The electrochemicalproton gradient is used to drive phos-phorylation of adenosine diphosphate(ADP) to adenosine triphosphate (ATP)through a molecular turbine linked to arotating catalytic machine. This storedwork potential is then used to drive ana-bolic (building up) processes in the syn-thetic direction by coupled reactions inwhich ATP is hydrolyzed to ADP andphosphate. At the metabolic level, someof the work available from spontaneouscatabolic (breaking down) reactions isalso used to drive ATP synthesis. The ATPis the “energy currency” of the cell. Theoverall reactions of metabolism, and ofphotosynthesis and respiration, can bebroken down into partial processes thatmatch the energy scale of the ATP/(ADP � phosphate) inter-conversion andthus facilitate the coupling between thesereactions through which some of the en-ergy is conserved for synthesis. By break-ing down the overall reactions to manypartial processes, and coupling thesethrough the ATP/ADP system, metabo-lism introduces a delay in entropic decay

of the intercepted energy, during whichthe available work can be used to drivethe biosphere.

The coupled reactions through whichphotosynthetic and respiratory electrontransfer chains drive ATP synthesis carrythe major flux of energy in the biosphere.Mitchell’s genius was to recognize that,contrary to earlier expectations, they areall transport processes, aspects of vecto-rial metabolism. The revolution in bioen-ergetics following from Mitchell’s ideas,introduced a spatiotemporal dimensioninto biochemistry that has brought intofocus the importance of membranes,transport, compartmentalization, etc.and that has colored the whole of biology[9].

INCORPORATING INFORMATIONCONTENT INTO THE ENERGYACCOUNTINGWhile in principle we can treat the bio-sphere as a physicochemical system op-erating under the standard laws, oneaspect of the energy balance presentsdifficulties. The problem is how to dealwith the informational content of thebiosphere. Information content comesinto the picture at two levels: that asso-ciated with the biochemistry of the bio-sphere and that associated with humancommunication. For the period of evo-lution before the invention of abstractrepresentations (about 30,000 yearsago), the energetic components couldbe subsumed under the chemical free-energy changes associated with synthe-sis of DNA. However, Homo sapiens as aspecies has learned how to communi-cate through language, to archive expe-rience, and to transmit information be-tween generations. This has led to asignificant change in the informationalcontent of the biosphere, or at least inits distribution. Apart from shifts in dis-tribution among life forms, no obviouscorresponding increase in the chemicalpotential attributable to the biospherehas been documented. There are dra-matic changes in the artifacts of civili-zation (cities, roads, ports, the Internet,etc.), and anthropogenic effects on thebiosphere itself, which might be takenas consequences of the increased so-phistication in exploitation of informa-

tion, and these can in principle bequantified through econometric meth-ods. But these are separate from theinformation content itself. How do weaccount in our energy balance for theinterventions of human society thathave led to such massive changes in theenvironment?

BIOLOGICAL INFORMATIONIt is common parlance when discussingthe ordered state of a system to refer toits “information content.” When used inthis context, an explicit connection be-tween information content and entropyis clearly established through the orga-nization of the system. Informationcontent is encoded in a physical systemthrough an ordering that requires inputof work, and the term is then used as asynonym for the increase in order (ne-gentropy) in its standard physical sense.The physicochemical changes in statecan be described through standardthermodynamics. Information is en-coded in the local gradients, physical orchemical, that define the state of eachcoding element, and the effectiveness ofstorage is directly related to the stabilityof the gradients formed. Attempts torecord the data in a fluid mediumwould lead to its disappearance by dif-fusion. This physicochemical usage re-lates to the physical state of the system,but tells us nothing about the semanticcontent of the information stored (themessage encoded). The term informa-tion content has alternatively been usedto cover this semantic aspect, althoughthe distinction between usages is oftenobscured. It is just this distinction be-tween the thermodynamic and seman-tic usages of the term that I wish todiscuss further here.

SEMANTIC CONTENT CANNOT BEEVALUATED BY PHYSICALMEASUREMENTIt is useful to recognize some interest-ing features of information content, anduse of this term in chemistry, physics,communication engineering, and biol-ogy. The first of these is quite generaland stems from the problem of relatinginformation content of a system to clas-sical thermodynamics. The classical


definition of entropy through probabil-ity, arising from the work of Carnot,Clausius, Boltzmann, and Gibbs, is re-lated to our general thermodynamicframework through its use in relation tophysical states and their temperature.The Boltzmann constant and the gasconstant are proportionality factors, thevalues of which can be determined ex-perimentally. They relate degrees offreedom in the system and its temper-ature to energy and determine theunits. Such constants allow informationcontent to be related to classical ther-modynamics, but only when specificallyrelated to physical states. However, itseems to me that a problem arises whenwe come to consider the semantic com-ponent. Although encoding and trans-mission of a message require changes inphysical states, the “meaning of themessage” itself is not related to thesephysical states in any thermodynamicsense. Of course, we need a “thermody-namic” framework (data storage, cod-ing, carrier, etc.), but we cannot tell byexamination of its physical propertieswhether a CD-ROM contains the com-plete works of Shakespeare or gibberish.The “meaning” is intangible; it occupiesno volume, has no mass, no electro-

magnetic property, and is not depen-dent on temperature. We can transferthe Encyclopedia Britannica to a harddisk in our car, but we have not identi-fied any mechanism through which wecould drive off using the “energy” re-leased as the meaning is unraveled. Wecan transmit the informational contentof the human genome at the speed oflight without altering the wavelength ofthe light. By use of appropriately tuneddevices, we can activate Voyager totransmit pictures by radio of the farplanets and receive the data, regardlessof the low temperature and near vac-uum of the intervening space. In theselast two examples, the energy cost oftransmitting a meaningful message orequivalent amounts of encoded ran-dom data would be the same; the datacan only be determined as meaningfulafter translation. The importance oftranslation can be highlighted by intro-duction of the notion that each of thesemessages would have been appropri-ately formatted and might have beendeeply encrypted. In each of these ex-amples, it is apparent that a message onwhich we place a high value behaves asif it had no energy content; it confers noadditional thermodynamic burden. Of

course, the information content atsome point has to be recorded in phys-ical format, and the archival medium isinterrogated and the transmission me-dium altered in the process with an en-ergy cost, and transmission of data re-quires a carrier, encoding and decodingfunctions, also with an energy cost; butthe value of the message itself, itsmeaning, is not a thermodynamic term.It depends on a translational context.

SHANNON’S RECOGNITION OF THEDISTINCTION BETWEEN ENGINEERINGAND SEMANTIC ASPECTSThe fidelity of transmission of informa-tion is the subject of a vast canon ofliterature on communication engineer-ing following from Shannon’s elegantdevelopment of Information Theory[12]. There has been a longstandingcontroversy1 over the relation between“informational entropy” and thermody-namic entropy, and much confusionarising from the loose usage of the terminformation. As noted above, it can beused in a context in which informa-tional entropy is equivalent to thermo-dynamic entropy, but only if related ex-plicitly to physical states. Alternatively,in the communications usage, it is ap-

1Shannon discussed the problem with VonNeumann [70], who advised him to callhis probability function “entropy”: “Youshould call it entropy, for two reasons. Inthe first place your uncertainty functionhas been used in statistical mechanics un-der that name, so it already has a name. Inthe second place, and more important, no-body knows what entropy really is, so in adebate you will always have the advan-tage.” Several physical chemists have ex-pressed the view that Shannon shouldnever have taken this joke seriously; asnoted in the text, the expressions are thesame only when they refer to the samephysical states. The modern usage of “in-formation” in physics seems more general;there is an implicit relation to entropythrough linkage to physical states, but theconstant is omitted, and the treatment isassumed to relate to binary states whichcan in principle by at any scale, down tosub-atomic spin states (the entropy is de-

termined by the total degrees of freedom ofmatter/energy). Shannon’s equation foruncertainty (H � �K ¥i�1

n pilog pi) andBoltzmann’s H-equation have the sameform, a probabilistic form first developedby DeMoivre in a treatise on games ofchance from the early 18th century (TheDoctrine of Chances, first published in1718). In Shannon’s words in relation toinformational entropy, “the constant Kmerely amounts to a choice of a unit ofmeasure.” If the unit of measure is directlyrelated to a physical system, kB, or its mo-lar equivalent, R, is appropriate. If twomessages are stored in an equivalent me-dium, use of the equation in a compara-tive context becomes quantitative, but notequivalent to a thermodynamic usage. If Krefers to an arbitrary coding element, com-parisons can be made quantitative if anydifference in coding elements is explicitlyrecognized (by use of two different Ks, bychoice of an arbitrary (usually binary)

coding element, or by omission of K). If abinary coding element is identified with abinary physical state, informational en-tropy has the thermodynamic meaningcommonly used in physics. However, thisusage introduces a problem of nomencla-ture; does “information” carry an implicitsemantic component? If so, then use ofinformation as synonymous with negent-ropy is misleading. If information is takento include the semantic component, thenit necessarily has to involve, at the least, animplicit receiver, with interpretationaland translational components of a physi-cal or biological apparatus.

In Weaver’s introduction to a later sum-mary of Shannon’s work [12, 71] hepointed out that the problem of infor-mation transfer can be partitioned intothree separate levels. Only the first ofthese is addressed explicitly by “classical”information theory: 1. How accuratelycan the symbols that encode the message


propriate when used comparatively,

with reference to general problems of

maintaining the integrity of the infor-

mation, for example, when the informa-

tional entropy, a probability function, is

being compared between two states

such as a transmitted and a received

message. The utility of informational

entropy lies in the fact that two encoded

messages, before and after transmis-

sion, for example, can be compared in

the context of the coding elements, even

when the encoding is different. If coding

is implicitly binary, and physical states

are treated as explicit binary elements,

informational entropy is again synony-

mous with thermodynamic entropy.

However, as Shannon pointed out in his

seminal paper [12], the semantic aspects

of information (the meaning of the mes-

sage) are irrelevant to such engineering

problems. Without examining the mes-

sage itself, we have no way of knowing its

semantic content, or its value. Implicit in

Shannon’s distinction is the question of

the thermodynamic status of semantic

content pointed out explicitly above.

We therefore appear to have three

different aspects of the information

content:

i. the thermodynamic side, in whichthe information content is related tothe ordering of systems in which thedata are stored, encoded, transmit-ted and restored,

ii. other engineering aspects associ-ated with the integrity of informa-tion coding, in which probabilisticmethods can be applied, and

iii. the semantic component of infor-mation transmission, where we lacka philosophical apparatus for relat-ing the information content to thethermodynamic world.

Clearly, the semantic content is ofimportance, but if, as Shannon seems toimply, we cannot measure an energycontent through its interaction with thephysical world, how can we bring it intothe context of measurement throughwhich we evaluate the energy balanceof the biosphere? There is an obverse ofthis question; can there be any messagewithout an encoding in some physicalform?

STATEMENT OF PURPOSEThe main thrust of this essay is to ex-plore these questions and their ramifi-

cations in biology and philosophy. I willtake the view, first, that the semanticcontent of a message cannot be evalu-ated in a conventional thermodynamicframework. The “meaning of the mes-sage” is not something which we canmeasure using the tools of science. Se-mantic content has a value that de-pends on a context and becomes appar-ent after processing through atranslational machinery. Second, I willtake it that use of the term “informa-tion” in a context that implies that se-mantic transmission can occur withouta thermodynamic vehicle, takes us out-side the scientific realm. By adhering tothese two principles, I will show thatsome interesting aspects of biology andphilosophy that have proved puzzling,and even intractable, can be opened upto examination, despite the limitationsimplicit in my first principle.

THE ENGINEERING PROBLEMS: THETHERMODYNAMIC COMPONENTThe engineering problem of data stor-age is straightforward; in order to storeinformation we have to modify a me-dium by “ordering” it locally, eitherthrough physical or chemical change.

be transmitted (‘the technical problem’)?2. How precisely do the transmitted sym-bols convey the desired meaning (‘thesemantics problem’)? 3. How effective isthe received message in changing con-duct (‘the effectiveness problem’)?Weaver has interesting ideas on each ofthese areas and recognizes the difficul-ties in formalizing any treatment of thelast two. Weinberger [72] has recentlydeveloped a treatment of the third themeby applying the Shannon equation to ananalysis of the consequences of action.He defines “pragmatic information” asthe information gain in the probabilitydistributions of the receiver’s actions,both before and after receipt of a mes-sage in some predefined ensemble. Thetreatment involves finding the mutualinformation between the ensemble ofmessages that could have been sent andthe ensemble of outcomes resulting fromthese messages. This allows for some in-teresting conclusions, but omission of

the constant K in the Shannon entropy ofmessage and outcome ensembles leadsone to wonder what states are scored inthe probability functions and how theycan be treated in real terms, as apposedto the abstract. In any evolving system,the pragmatic solution is of course sim-pler and more savage: survival of thefittest.

Weinberger’s treatment does not ad-dress the status of semantic content—thecomponent of Weaver’s partitioning pre-viously recognized by Shannon—thoughhe does refer to attempts to deal with itin the context of complexity (see refer-ences below). However, he offers theopinion that none of these treatmentshave been very successful, and it seemsunlikely to me that any treatment thatfails to recognize that semantic contenthas value only in context, and only aftertranslation will be successful. An un-abashed recognition that semantic con-tent contributes no additional thermo-

dynamic burden does seem to lead to a

productive line of thought. However, I

am aware that recognition of the impor-

tance of a component of biology that is

not accessible to physical measurement

will be anathema to many of my scien-

tific colleagues. I should therefore make

it clear (again) that I am not proposing

anything unnatural here. As long as the

message is tied to thermodynamic repre-

sentation, we can deal with it in a secure

framework of realistic philosophy, and

pursue the interesting implications.

For a discussion of different perspec-

tives on the relation between classical

thermodynamics and information the-

ory; see [73–76]. For other references see

Brian Harper at (http://www.asa3.org/

archive/evolution/199707/0124.html).

Further discussion of complexity in the

semantic context can be found in [77–

80].


This necessarily involves an input ofwork. A particular message is storedthrough an ordered set of codons, orthrough some equivalent change in themedium.2 Does this necessarily meanthat the gradients are exploited in trans-mission of information? In principle,the work involved in defining a codoncould be returned by coupling to anappropriate system in the surround-ings. This would require a device inwhich work-terms, scale, and patternwere appropriately matched. But trans-mission of information using the storedwork would necessarily lead to someloss of the stored information; it mightbe useful for a one-shot transmission,but not for archival purposes. Any read-ing would deplete the stored informa-tion; eventually all information wouldbe lost from the original medium, es-sentially by diffusion. In practice, in allsystems that depend on fidelity andlong-term storage, information is readfrom a medium by interrogation of thestate with an input of work, rather thanthrough output by exploitation of thework potential stored.

The same general engineering prob-lems are dealt with in biological sys-tems. Storage of genetic information ina stable form requires work throughsynthesis of DNA polymers, and the se-mantic content is abstracted by specificchemical interrogation, again with in-put of work. The data are stored in lin-ear format, information is encoded inthe sequence, and translation leads ei-ther to replication or to synthesis of lin-ear molecules in which the code is di-rectly reflected. Biological diversityresults from exploitation of the combi-natorial possibilities of the code. “Cod-ing elements” are recognizable by se-

quence characteristics encoding proteinsequences. The triplet code of differentcombinations in sequence of the 4bases, allows coding for 20 amino acids,with excess informational capacity forredundancy, punctuation, and control.The combinatorial possibilities pro-vided by 20 amino acids with differentphysicochemical properties, each withseveral conformers, allow for a muchlarger range of proteins with differentcatalytic, control, and structural func-tions. Many of the so-called “noncod-ing” elements encode sequence ex-pressed as RNA polymers that haveroles in the translational machinery,control and timing. All these transla-tional products in turn provide a basisfor the higher orders of combinatorialcomplexity that lead to the full range ofdiversity of the biosphere.

SEMANTIC CONTENT HAS VALUEONLY IN AN APPROPRIATE CONTEXTIn order to bring out the distinction be-tween the engineering and semanticuses of the term “information content,”I will introduce the next part of my dis-cussion through Olsen’s “miracle.” Mycolleague Gary Olsen told me how inhigh school he had undertaken a phys-ics exercise in which he and his fellowstudents were asked to toss a coin andrecord the statistics. Being a conscien-tious student, Gary recorded not onlythe net yield of heads and tails, but alsothe sequence, for a thousand or so tri-als. Not surprisingly, he found that thetwo options occurred with about equalprobability. However, Gary noted thatthe probability of obtaining the samesequence of heads and tails in any sub-sequent trial was exceedingly low,hence the miracle. The important les-son from this example is that the se-mantic content of the sequence had aspecial significance, but only in relationto its peculiar context. There was no lessinformational entropy in this particularsequence than in any of the otherequally improbable sequences exploredby the class, but it became significantbecause Gary noticed its interesting se-mantic properties.

We will return to this distinction be-tween a statistical and a sequential view

later. For the moment, I want to empha-size that the value of this particularmessage depended on context. Thisbrings up the interesting possibility thatthe semantic content of all biologicalinformation must fall into this samegeneral contextual framework; it hasvalue only in relation to a biologicalcontext. The idea that semantic contentdoes not contribute to the thermody-namic cost of communication mightseem to take consideration of this topicoutside the domain of science. How-ever, if semantic content is tied to athermodynamic representation, thenthe latter can provide constraints thatjustify further consideration. In the fol-lowing, I will explore the consequencesof accepting this proposition.

THE DISTINCTION BETWEENTHERMODYNAMIC AND SEMANTICCOMPONENTS OF GENETICINHERITANCEAs noted above, the engineering side ofbiological semantic transmission ishandled at the level of coding and in-formation storage in DNA (or RNA).However, the semantic content of a par-ticular encoding in the DNA sequenceof a genome, the genotype, has signifi-cance only after translation, and expres-sion in some functional context, such asa particular protein, control mecha-nism, behavior, organism, etc., featureslumped together as the phenotype. Not-withstanding the dominant role of thegene [13], it is the phenotype that istested by evolution. Dawkins [14] sug-gests that we “. . . think of phenotypesas the levers of power by which the suc-cessful replicators manipulate their wayinto the next generation . . .”. The dis-tinction between the gene as a physicalentity, and the message carried by thegene, is subtle. Although the metaboliccost of synthesis of DNA can be quan-tified, the cost is sequence-indepen-dent. The informational entropy of aDNA sequence could be scored througha probability function, and probabilityapproaches are of importance in com-parison of sequences and in determina-tion of the history and fidelity (or oth-erwise) of information transmissionwithin and between families, but the

2Writing is an interesting case: a blankmedium is modified by a local coveringof opaque dye, using symbols that origi-nated as pictograms, but have developedmore abstract forms. The data encodedcan be read through use of light reflectedfrom the surface and focused onto anappropriate decoding device that deci-phers the symbols, most commonly theeye/brain interface.


probability score for a particular se-quence has no greater significance thanthat for many other of the similarly im-probable instances derived from thesame set of codons. Of course, in a com-parative context, the informational en-tropy would have value in scoring thesimilarity of two sequences. It is themessage, however, that is unique. Theparticular sequence of a particular ge-nome defines a particular semanticcontent, the message that is passed on,and hence the particular organism, and,with small local variations, the species.It is the message that has a value, butonly in the context of Darwinian evolu-tion (with all the necessary complexitiesin terms of mechanism).

Interestingly, the distinction be-tween semantic and mechanistic aspectof information transmission was al-ready noted by Gamow in the days ofthe RNA tie club, shortly after the firstintroduction of the Watson-Crick model[15]. He pointed out that the problem ofsemantic transmission in inheritance isseparate from the precise mechanism, adistinction implicit in Shannon’s earliertreatment of information theory. Themechanism that has emerged frommuch subsequent research, although atfirst sight much less elegant than sev-eral of those proposed, is optimal andalmost universal, but not accidental[16]. However, the mechanism is part ofthe engineering side. A reading of thestored information does not change thephysical state of the storage medium orthe information stored. In contrast, themeaning (the semantic content) is ap-parent only after translation, and itsthermodynamic status is tenuous. Thevalue of the message is tested by evolu-tion. The organism is fitted through itsinheritance of information for a partic-ular ecological niche in which it has thebest chance of survival. The niche isitself a complex matrix of geological, bi-ological, and spatiotemporal interac-tions in which a dominant term is com-petition or collaboration within andbetween species in the quest for limitedresources. But each of the species inter-acting in any particular ecological con-text is itself defined by its own inheritedinformation, which determines the be-

havioral repertoire, and hence the pa-rameters of interaction. The informa-tional context, the multiple feedbackloops between different species, istherefore extremely complex. Some fla-vor of this is given by Darwin’s elegantclosing remarks in The Origin of Species

It is interesting to contemplate anentangled bank, clothed withmany plants of many kinds, withbirds singing on the bushes, withvarious insects flitting about, andwith worms crawling through thedamp earth, and to reflect thatthese elaborately constructedforms, so different from eachother, and dependent on eachother in so complex a manner,have all been produced by lawsacting around us . . .. There isgrandeur in this view of life. . . from so simple a beginning

endless forms most beautiful andmost wonderful have been, andare being, evolved.

It might be argued that, with the rev-olution in biology following theWatson-Crick model, we know a greatdeal about the genetic message; afterall, we can read the code for Craig Ven-ter, warts and all, and there’s nothinginsubstantial about it! However, our re-cent acquisition of an ability to decodethe genetic message is a reflection ofour understanding of the first level ofdecoding, leading to translation intoprotein. In the protocols for analysis ofgenes, which are mainly aimed to the“coding elements,” we have developedan abstract analogue of the naturaltranslational processing at this level,roughly equivalent to optical characterrecognition compared to an apprecia-tion of Shakespeare. We are far frombeing able to determine from Venter’sDNA what makes him a successful hu-

man being. If this higher level of sophis-tication is ever realized, it will be foundstill to be an analogue of translationalprocessing. However, it will require adeeper knowledge of biology than iscurrently available; details of the devel-opmental environment from concep-tion to the time of his maturation. Thiswould have to include a complete as-sessment of metabolic and cellular “fit-ness” from the molecular level on up,and a complete analysis of his personal“ecology,” including not only the phys-ical, but also mental, psychological, andsocial development, culminating in theniche to which he is fitted. A successfulanalysis might conceivably, in the fu-ture, be realized in retrospective, but, asthe discussion on emergence below willshow, not in predictive mode.

A point needs to be emphasized inrelation to the distinction above, whichrefers to information transmission be-tween generations. The synthesis ofDNA leading to the germ cell line, orany other reproducing state, is clearly acopying function. The metabolic cost ispart of the engineering side and is arelatively small fraction of the syntheticand maintenance costs of the cell. Incontrast, the semantic content is de-fined by the message transmitted be-tween generations. This depends ontransfer of the message carrier into atranslationally competent environment.In bacteria, and in asexual reproductionin the protista, simple cell divisiontransfers a replica of the genome at thesame time as it splits off a chunk of thecellular machinery to provide the trans-lational function. In sexual reproduc-tion in higher organisms, the haploidgametes are highly specialized cells,with motile (male) and more sessile (fe-male) forms. The female gamete carriesthe somatic machinery that providesthe translational apparatus. The majorbulk of the mature phenotype, gener-ated by asexual (mitotic) reproductionand tissue specialization, plays only amechanical role. In all cases, the se-mantic content changes slowly by es-sentially stochastic processes—muta-tion and selection— over evolutionarytime. If this had any energy cost overand above that of copying, it would

It is the message, however, thatis unique. The particular

sequence of a particular genomedefines a particular semantic

content, . . .


have been amortized over the evolu-tionary span. Assuming a common ori-gin, this timescale, about 4 billion years,is the same for all extant organisms.

THE INTERPLAY BETWEEN THEBIOSPHERE AND THE PHYSICALENVIRONMENTAs a comparison of a global map ofpopulation density with one showingthe environmental determinants ofphotosynthetic productivity will dem-onstrate, the biosphere is highly depen-dent on the physical environment. Inaddition to temperature, water avail-ability, nutrients, and insolation, thedistribution of life depends on such fac-tors as the latent heat effects mentionedearlier. The winds, currents, and tidesdistribute thermal energy around theglobe, ameliorating the seasonal changein insolation, and life on land dependson the fresh water delivered by the hy-drological cycle. We should also notethat the biosphere has, in turn, affectedthe environment. Among other things,life has changed the energy distributionof the light absorbed. The most obviouseffect has been through the color ofplants. The earth looks green fromspace; it is this that allows the seasonalsweep of spring from hemisphere tohemisphere to be monitored by satellite[5]. Another major effect has been onthe atmosphere, which is in a very dif-ferent state from that which would befound without life. In particular, theabundance of oxygen and ozone andthe rather low concentration of CO2 area direct consequence of the invention ofoxygenic photosynthesis and thechange from an anaerobic to mainlyaerobic environment some 2.2 billionyears ago. Other less obvious effectshave followed; the whole redox poise ofthe atmosphere, oceans, and surfacechemistry has been shifted with theevolution of an aerobic atmosphere.Again, the web of interactions has beenquite complex. The modern biosphereis, of course, adapted to the present en-vironment, and most of its characteris-tic features are dependent on it. Al-though the biosphere is relativelyrobust to local disasters, and adaptswell to long-term changes, the limited

information we have about response toglobal changes occurring over short pe-riods points to a more fragile balance.The evolutionary record shows numer-ous “extinction events,” only a few ofwhich have been satisfactorily ex-plained. This might give us pause incontemplating the more dramatic an-thropogenic changes we have imposedon the environment in recent years.

It takes no great stretch of imagina-tion to see from the above argumentsthat the biosphere has evolved as awhole, and in a dynamic exchange withthe physical environment. Interactionsbetween organisms feed back on thegenotype of each contributing speciesthrough survival of the fittest. The se-mantic content of the biosphere, as re-flected in the sum of the semantic con-tents of all genotypes, has beenchanging so as to adapt it to the oppor-tunities available. Recently, those adapta-tions have included the peculiar abilitiesof our own species to expand the infor-mational context on the semantic side.However, the starting point for each newgeneration is the information contentthat has survived from previous changes.

SEMANTIC TRANSMISSION IN HUMANCOMMUNICATION AND THE“INSUBSTANTIAL” IN LITERATUREAND PHILOSOPHYThe general conclusion that the seman-tic significance of information is deter-mined by the context seems to applyequally well to the sophisticated end ofthe informational spectrum and mightbe more easily appreciated there. Theworks of Shakespeare (or Einstein’sviews on relativity) are of significanceonly in the context of an informed au-dience, equipped to translate both thecode (language) and the semantic con-tent (meaning). The improbability ofsuch works is of course remarkable. It

might take a zillion monkeys the life-time of the universe to generate theworks of Shakespeare by typing at ran-dom. But so what? Any combination ofthe same alphanumeric and formattingsymbols could have similar informa-tional entropy. Each one would likely beunique. We value only the one, and thatevaluation is specific to our particularcultural, linguistic, and temporal context.

An instructive experiment you cantry is to ask a friend to close her eyes,and then say some phrase like “We aresuch stuff as dreams are made on, andour little life is rounded with a sleep.”You can then ask what came into herhead when she heard these words. Youcould ask yourself the same questionnow, after reading the phrase. You willget a wide range of answers—“What’sthis guy up to here?,” “Isn’t that a quotefrom somewhere?,” “That’s an interest-ing proposition,” or maybe even “Why,that’s Prospero’s speech from Act IV ofthe Tempest”! All these are wrong. Infact, nothing enters the head undersuch circumstance. No matter whatchannel is used (sound or sight here),the medium for transmission is physicalor chemical and is intercepted before itenters the head (Figure 2). What is per-ceived as “coming into your head” in-volves a hierarchical set of translationaland interpretational processing func-tions, triggering associative memoriesalready resident in the mind. Thephrase has meaning only because wehave a mind and an education and thenecessary translational machinery.

The insubstantial nature of mentalprocesses has been a recurring theme inpoetry and literature, as the quotationabove will illustrate. The same theme re-curs in philosophy, going back at least toParmenides, Platonic Forms, and Aristo-tle’s phantasma, popping up in mediae-val theology in debates about universalsand surfacing again more recently, for ex-ample, in Descartes’ distinction betweenmind and body, in Bishop Berkeley’s “in-genious sophistry,” in the Kantian dis-tinction between noumena and phenom-ena, and in Putnam’s brain-in-a-vatscenarios (borrowed in the Matrix mov-ies), and is discussed more extensivelybelow. The recognition of two contribu-

It takes no great stretch ofimagination to see from the abovearguments that the biosphere has

evolved as a whole, and in adynamic exchange with the

physical environment.


tions to information content and my dis-cussion of the thermodynamic inconse-quence of semantic content both speakto this recurring theme. If the value ofsemantic content depends on context,then it also depends on the translationalmachinery through which that content isappreciated. For that component trans-mitted through DNA, the translationalmachinery is the common biochemistryof the biosphere. For the semantic con-tent of our civilization, however, the ma-chinery is that of the individual mind.Further pursuit of this theme in the con-text of the mind requires that we addressthe question of the origin of this remark-able contraption.

BIOLOGICAL COMPLEXITY, BEHAVIOR,AND THE PERCEPTION OF TIMEThe evolutionary context in which theabove discussion is framed introduces atemporal element, the �4 billion years ofdevelopment of the present biosphere,and the recent evolution of humankind(Richard Dawkins has a nice “reverse” ac-count in The Ancestor’s Tale [17]). Ourdiscussion of the semantic content of thegenome requires consideration of time,the perception of time, and the evolution-ary perspective, because the content hasvalue only in a dynamic context that in-cludes behavior among other temporal

aspects. It is interesting to enquire when

time became significant for the inhabit-

ants of the biosphere.

All living things show behavior; they

respond to stimuli from the environ-

ment. The environment is dynamic, and

the behavior is adaptive. Because of

this, the behavior has a temporal com-

ponent; the organism anticipates its fu-

ture in relation to the present. The suc-

cess of this behavioral anticipation is

essential to survival. In this sense, all

living organisms are aware of time. To

sidestep the accusation of an anthropo-

centric view, I will introduce as a term

for this perception of time, the word

chronognosis,3 with the adjectival form

chronognostic. This avoids the use of

words like awareness and perception,

which carry implications of a higher or-

der of mental activity.

Bacteria have relatively simple pat-

terns of behavior; the organism re-

sponds to gradients, either physical or

of chemical effectors, by modifying the

direction of rotation of a bundle of fla-

gella. This determines whether they ei-

ther swim in a straight line (clockwise

rotation: the flagella are all in phase), or

tumble (counter clockwise rotation: the

flagella fly apart), in order to muddle

their way to (or from) the source of at-

tractants (or repellants).4 A change in

behavior is triggered by a simple “per-

ception machine.” For many bacteria,

the behavior indicates a relatively short

range of temporal perception (the chro-

nognostic range) determined by a well-

characterized set of mechanisms. These

include detection of attractant or repel-

lant molecules by binding to chemo- (or

photo-) receptor proteins that span the

cell membrane, followed by signal

transmission through a conformational

change that transmits a message from

outside to inside. This activates a “cas-

cade” involving various pathways, typi-

3From Greek �� (chronos, time) �

�� (gnosis, investigation, knowledgegained from experience). An alternativeword, originally introduced in the contextof bee behavior but now used more exten-sively, might be zeitgedachtnis, Germanfor time-memory [81], but this also impliesa higher-order memory function.

4See the References [82– 85] for articleson biophysical, biochemical, and physi-ological aspects of the flagella motor andbacterial motility.

FIGURE 2

The mind, its external interfaces, and the thermodynamic carriers for sensory and semantic input. Penfield’s sensory homunculus, projected on the cortex ofthe brain (left), or as a model distorted in proportion to the area of the cortex concerned with touch (middle), and the physical or chemical vehicles that impingeon different sensory interfaces (right). (The model is in the Natural History Museum, London, and has been released to the public domain. Models for other sensescould be constructed along similar lines). [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]


cally a two-component histidine ki-nase/response regulator system thatleads to modulation of flagella re-sponse. An essential component of themechanism is adaptation, a biochem-ical process with variable time con-stants that play off against each otherto allow detection of a change in gra-dient (Figure 3). The precise role var-ies between organisms, but the systemcan be regarded as involving compet-

ing decay and refresh pathways, inwhich the time constants for somefunctions are under control by thestate of activation of the receptor. Thisallows recognition of changes in gra-dients, with a time scale on the sec-onds-to-minutes range. We might takethis range of time constant as definingtheir chronognostic range. However,for a worm, an ant, a fish, or a dog, therange is obviously longer. With the in-

creasing complexity of metazoans, de-velopment of a nervous system, differ-entiation of organs of sense anddevelopment of the head-tail polarity,and a brain and memory in higher ani-mals, the scope for behavioral complex-ity also increases, and along with this,the complexity of mechanisms for per-ception and the chronognostic range.

The temporal context of behavior isnot restricted to animals. Mobility is the

FIGURE 3

Dissecting the biochemical machinery underlying bacterial chronognosis. The apparatus through which the bacterium detects changes in gradients includes thefollowing elements. Detection of attractant or repellant molecules through a chemoreceptor (MCP), which transmits a signal across the cell membrane; thesubsequent triggering of a signal transmission cascade through a two-component histidine kinase (CheA) and response regulator (CheY) system; and modulationof flagella response (insert, lower left). This involves phosphorylation of CheY (catalyzed by CheA) or dephosphorylation (CheC, CheZ) to change the directionof rotation. An essential component of the mechanism is adaptation, a biochemical process with variable time constants involving methylation (catalyzed byCheR), and demethylation (catalyzed by CheB) by reversible esterification of acidic side chains in the receptor, which modulate the degree of aggregation, likelyby neutralizing the coulombic repulsion from like charges. This leads to deactivation or activation of the signal transmission pathway. The precise role variesbetween species, but the system can be regarded as involving competing decay and refresh pathways, in which the time constants for some functions are undercontrol by the state of activation of the receptor. This allows recognition of gradients—changes in local concentration—with a time scale on theseconds-to-minutes range. Details of mechanism vary in different bacterial families; the scheme here is appropriate for Bacillus subtilis, in which an additionalmechanism for amidation and deamidation (CheD) of the acidic residues also contributes. An interesting variant recently reported [99, 100] is a activation ofa similar CheA kinase through binding of sensory rhodopsin II to its membrane partner HtrII on photoactivation in haloarchaea. Art work adapted from Ref. 85by addition of a serine receptor dimer model from Ref. 84.


main feature distinguishing animals

from plants and introduces a spatial el-

ement. I will call this the chronognostic

scope, given by the product of spatial

range and temporal perception. We can

recognize that plants have a much more

limited scope than animals. However,

although the time scale of response and

their mobility make behavior more ob-

vious in animals, the need for regula-

tion of the photosynthetic apparatus in

response to the diurnal cycle requires

anticipatory behavior in even the sim-

plest photosynthetic organisms. At the

lowest level, the cyanobacteria use a cir-

cadian biochemical clock, whose pro-

teins are encoded by the kai genes,5

which extends their chronognostic

range to several days. In higher plants,

the daily and seasonal responses in an-

ticipation of sunlight, temperature, wa-

ter, etc. demonstrate a range at least of

years and are regulated by more com-

plex clocks, allowing quite sophisticated

responses sometimes referred to as

plant intelligence [18].

The circadian clocks were rein-

vented in the animal world,6 and have

given all complex animals built-in, bio-

chemical, temporal reference systems

that regulate their time-dependent me-

tabolism and physiology [19, 20]. It

seems likely that all living things have

built-in biochemical timers, directly

linked thereby to the semantic inheri-

tance through DNA and that these are

also important as determinants in be-

havior. The primitive timers might be

thought of as the first steps towards per-

ception.

THE EMERGENCE OF BIOLOGICALCOMPLEXITYThe increasing complexity of livingthings has depended on the evolutionof different ways of exploiting the com-binatorial possibilities inherited in theinformation encoded in the genotype.From the above discussion it will beobvious that semantic inheritance is anecessary corollary of life. Any self-rep-licating system requires a “genetic”message—the instructions for replica-tion and for the phenotype—with itsnecessary engineering and semanticcomponents. While recognizing thatboth are essential, it is of interest tonote that the semantic content hasmany of the properties of the elan vital[21], except that, although it may bevital, it is not a force. The concept of a“vital force” has invited the scorn of thescientific community [22], so, in draw-ing this analogy, it would be as well tore-emphasize that the thermodynamiccomponent ties semantic transmissionto the physical world. Nevertheless, it isthe message that is transmitted be-tween generations, and this nonther-modynamic component is essential toall life. For all extant forms, the seman-tic content has been honed throughevolution; any current message repre-sents the result of an investigation ofthe nature of reality through the ulti-mate test of survival.

As noted many years ago by Eigen[23–25], evolution is a continuing explo-ration of the possibilities opened up bycombinatorial complexity. In this sec-tion, I want to show that it is only inevolving systems, based on transmis-

sion, translation, and mutation of a se-mantic heritage and that this potentialfor higher complexity can be practicallyrealized. Paul Davies has suggested that“Physical systems are causally openwhen they exceed a certain level ofcomplexity— determined by the infor-mation-processing capacity of the uni-verse,” which is taken to be in the range10120–10122 bits (see Davies [26] and ref-erences therein). There are enough bitsto specify the configurational possibili-ties for a protein of only 60 –90 aminoacids, or the coding potential in �200nucleotides, hardly enough for a Lapla-cian intelligence (or any other designer;Figure 4) to use as a deterministic basisfor the complexity of the biosphere. Da-vies therefore concludes that life is anemergent phenomenon. A similartheme was introduced much earlier byLevinthal in discussion of protein fold-ing pathways [27]. The informationalcomplexity of the genome is deter-mined by the 4n possibilities inherent ina genetic code of four bases in an n-length sequence. For translation to pro-tein, the triplet code allows specifica-tion of the sequences made up from the20 amino acids (each protein of definedlength), with redundant coding, punc-tuation, and control built into the ex-cess capacity. The combinatorial possi-bilities arising from 20 amino acids withdifferent physicochemical properties,each of which can be deployed in sev-eral different conformers, provide awide range of opportunities for proteinconfiguration [assuming 5 conformerswe have (20 _ 5)60 for a protein of 60amino acids, to match Davies’ 10120 in-

5While some elements of the kai systemseem to be restricted to the cyanobac-teria, other kai genes are more widelyspread. Homologs of the kaiC gene,encoding the RecA-superfamily ofATPases implicated in signal transduc-tion, are found throughout the bacte-ria and archaea and are involved incircadian functions in some of these[86].6The cryptochrome photoreceptors thatprovide synchronization of circadianclocks through photoactivation by day-

light are found throughout the three do-mains of life. The period, cycle, andtimeless proteins of Drosophila and theclock proteins of vertebrates, each sug-gested as having a circadian function,seem to be fairly widespread. Most workreported has been on insects and verte-brates. Period 1 homologs are found inmollusks and nematodes, as well as inthe higher animal families. Period 2 ho-mologs are found in bacteria and manymetazoans, but not necessary in a circa-dian context. Timeless homologs are

found throughout the Eukarya, mostlyidentified in Metazoa, including plantsand fungi, but are rare in bacteria. Clockprotein homologs with timing functionare in all vertebrates, and also in insects.Clock homologs of unknown functionare more widely spread, from Dictyoste-lium onward, in animal, plant andfungal branches of the Eukaryota. Ofinterest to anyone working in bioener-getics, clk1 encodes a di-iron hydroxy-lase in the ubiquinone synthesis path-way [87, 88].


formation bits], exploited in the manystructural, enzymatic, and control rolesfound in living organisms.

It should be noted in the context ofprotein synthesis that the translationfrom a linear DNA code [essentially aone- or two-dimensional (2D) informa-tional system] to proteins does not in-volve any increase in information con-tent, rather the reverse. Apart from thematernally inherited mitochondrialDNA in eukaryotes, and the initial so-matic machinery [8, 9] (not an unim-portant component, since it providesthe essential translational apparatus),the genome necessarily contains all theinformation for specification of thephenotype. For any particular span, the

information content is related to thelength of the coding sequence and thenumber of coding states. The 3-to-1length ratio implicit in the use of 3 DNAcodons to specify each amino acid, andthe ratio of 4 to 20 in coding elementsplay off against each other. The infor-mation content in DNA is richer thanthat of the expressed protein, because italso includes redundancy, syntax, con-trol, and other functions associatedwith the so-called noncoding elements.The translational processing involves adegradation of information content, butan increase in complexity. How is thisachieved? The trick is to increase thecombinatorial possibilities, and themechanism works through the transla-tion to different levels of complexity ex-pressed in the third and fourth dimen-sions (Figure 5). The topology ofproteins involves local specificities thatarise from the different properties ofamino acid side chains, and the combi-nation of forces in three dimensions toprovide spatially unique physicochemi-cal surfaces. These are available, for ex-ample, as external recognition tem-plates, as cavities tailored for specificbinding of substrates associated withcatalysis or control, or as surfaces forpartitioning between lipidic and aque-ous phases, essential to the functional-ity of cell membranes and their vecto-rial metabolism. Genomes for humans,and for many other higher animals andplants, encode some 25–30,000 proteinsor subunits, most of which have lengthgreater than Davies’ “emergence limit.”Interactions between these, and withco-opted molecules (lipids, polysaccha-rides, nucleic acids, metabolites, coen-zymes, prosthetic groups, etc.) involvedin metabolism, construction, recogni-tion, and function of cells and tissues,extend the combinatorial possibilitiesto a wider range of phenotypical com-plexity. The emergent property of life istherefore clearly dependent on the se-mantic processing. Essential prerequi-

sites for the amplification in complexityare the translational steps throughwhich the genetic message is inter-preted, initially in protein. These pro-vide an increase in combinatorial op-portunities through the interactionspecificities and catalytic potential builtinto the protein structures. The varia-tion in the properties of the specific sur-faces through mutation and the selec-tion of useful variants by evolutionallow, over time, for an exploration ofmany different structures.

That selection is an essential compo-nent, even at the initial level, becomesobvious if we invert the problem. Inprinciple (and in practice for small pro-teins) we could synthesize DNA to en-code a designer protein, and generate itusing the cellular machinery abstractedin vitro, or by borrowing an in vivo ma-chinery. If we wanted to fully realize thedesign potential, we would come upagainst the same problem as Levinthal’sparadox or Davies’ Laplacian intelli-gence: we would not have the compu-tational resources to do a comprehen-sive job. Evolution does the job withoutdesign. A large number of configura-tions are explored over time simply bymaking mistakes. The evolutionary pro-cess discards the duds.

The interaction specificities arisingfrom translation to structures with 3Dsurfaces represent only one level of po-tential combinatorial possibility. Thefourth dimension comes in at several lev-els. The stochastic diffusional processesthrough which molecules bump intoeach other are tied simply to time, butreaction mechanisms and catalysis arerelated to time through more complexfunctions. Enzymatic catalysis allowsmore turnovers per unit time (typically byfactors in the range 1012), but also allowsfor specificity in reaction steering, and forcontrol, and hence for the evolution ofcomplex metabolic webs. The timing ofontogeny, and of the specification for de-velopment of tissues and organs in morecomplex organisms, is another level, asyet poorly understood. Tissue specifici-ties allow for division of labor and theefficient coordination of complex higherfunctions. The selective role of evolutionrequires interactions at the level of ecol-

FIGURE 4

Pierre Simon Laplace (1749–1827) and thedeterministic view: “We ought to regard thepresent state of the universe as the effect ofthe antecedent state and as the cause ofthe state to follow. An intelligence knowingall the forces acting in nature at a giveninstance, as well as the momentary posi-tions of all things in the universe, would beable to comprehend the motions of thelargest bodies as well as the lightest atomsin the world, provided that its intellect weresufficiently powerful to subject all data toanalysis; to it nothing would be uncertain,the future as well as the past would bepresent to its eyes . . .”. Laplace, 1820, Aphilosophical essay on probabilities. [Colorfigure can be viewed in the online issue,which is available at www.interscience.wiley.com]

The emergent property of life istherefore clearly dependent on

the semantic processing.


ogy and behavior, so that combinatorialcomplexity is extended even further intothe temporal dimension. The combinato-rial potential will obvious increases withtime, but defining a function to describequite how presents a challenge; likely,some power law dependence on dimen-sionality will emerge.

An analogy might be useful here. Thedigital encoding on a CD of Beethoven’sFifth or a reproduction of the Mona Lisaor the human genome as text wouldlook boringly similar if examined beforetranslation—just a string of zeros andones. Of course the encoding containsthe necessary information and format-

ting, but it is only after your computerrenders the file through a translationalprocessing that something of beauty orwonder is revealed. Not that the com-puter cares—many additional levels oftranslation are required before we canappreciate these facets.

A SIMPLE RNA WORLD HAS LIMITEDPOTENTIAL FOR COMPLEXITYThe hypothesis outlined above to ac-count for the emergent properties of lifehas implications for early evolution. In-formation transfer through DNA andtranslational expression in a protein-aceous phenotype follow a mechanisticpattern that is common to all extantforms. This is interpreted as evidencefor a common ancestral condition thatevolved some 3.5 billion years ago toallow the development of the modernbiosphere [28, 29]. Since some transla-tional step is essential for exploration ofcombinatorial complexity, a criticalstage in evolution must have been theintroduction of this initial amplificationprocess. A simple “RNA world” wouldhave had the same level of complexityfor the genotype and phenotype. TheH-bond matching between the basesthat make up RNA allows a coding po-tential, but structural flexibility of RNAallows more degrees of freedom thanDNA. In extant forms, these two prop-erties are exploited in recognition dur-ing translational processing at t-RNAand m-RNA levels, in control functions,in specific catalysis by ribozymes, andpossibly in timing. An RNA world wouldtherefore have had greater opportuni-ties for complexity than one based onthe structurally constrained DNA. In themechanism now used, DNA serves onlyfor the genotype; the translational stepsleading to the phenotype, starting withprotein synthesis, provide the pathwayto complexity. Whatever possibilitiesthe multiple roles of RNA might havepresented for a separate evolutionarymode, they have not survived as themainstream. The coding potential is ex-ploited in some viruses, but these de-pend on fooling the translational ma-chinery of the host. In other life forms,the functional possibilities are exploitedonly in ancillary roles. If an RNA world

FIGURE 5

Top: Combinatorial complexity on going from one to four dimensions. Left, the double helix containstwo complementary copies of the genome with identical information content. In bacteria, this singlegenome restricts dimensionality to 1D, but in sexual reproduction additional combinatorial levels areopened, effectively in 2D, by chromosomal crossover. The 3D level (middle) is represented by the bc1

complex catalytic core (PDB ID 2bcc, 3bcc) [101], with the cytochrome b subunit surface colored toshow its physicochemical surface (as electrostatic potential: red, negative; blue, positive). Thiscomplex has two catalytic sites for oxidation or reduction of ubiquinone species (or binding of inhibitorymimics), internal pockets for binding of four metal centers (3 hemes and an iron-sulfur cluster), two externalcatalytic interface for interaction with a mobile domain of the iron-sulfur protein, a hydrophobic membraneinterface, several interfaces for interaction with other subunits, and a dimeric interface. Bottom: The adeninenucleotide translocator (PBD ID 1okc). The surface is colored to show electrostatic potential. This stereo view(for crossed-eye viewing) allows a glimpse into the pocket where adenine nucleotide binds, which isoccupied here by the inhibitor carboxyatractyloside. This protein is thought to act in dimeric form toexchange ADP for ATP� across the mitochondrial membrane through large conformational changes; thesurface is quite complex and shows not only the internal binding pocket, but also two aqueous polarinterfaces, a hydrophobic collar that interfaces with the membrane lipid, and a dimeric interface.


was precursor to the DNA-proteinworld, it will be necessary to map out apathway through which the essentialtranslational step was introduced.

HUMAN INTELLIGENCE AND THEAWARENESS OF EXTENDED TIMEThe interesting correlation between com-plexity achieved—the depth of exploita-tion of the semantic content—and thedepth of perception of time has been al-luded to above. The increase in chronog-nostic range allows a power law increasein the combinatorial possibilities throughpermutations in the temporal dimension.Until �100,000 years ago, chronognosticrange at the species level was linked di-rectly to behavioral complexity and henceto information encoded at the DNA level.The chronognostic range of human kindwas likely not greatly in excess of that ofother higher mammals; orally transmit-ted traditions become myths after a fewgenerations. However, all this changedwith the evolution of sophisticated lan-guage and abstract representation. It isclear that with the invention of archivalforms of information and transmissionbetween generations, inherited semanticcontent was no longer restricted to thatencoded in DNA. With the invention ofnew mechanisms for transfer of informa-tion, the chronognostic range has also in-creased, in both temporal directions. Thegrowth has expanded even more dramat-ically as civilization has enabled the insti-tutions, both physical and philosophical,through which transmission and storageof information has been formalized. Notleast among these have been the socialand economic structures that have liber-ated a small fraction of its more privilegedmembers for the task of exploring andextolling the human condition.

ENTROPY AND THE ARROW OF TIMEEddington coined the phase “arrow oftime” to describe an implicit direc-tionality in time. This was based onBoltzmann’s recognition that for anisolated system the spontaneous di-rection of any process was toward in-creased entropy or disorder. Althoughseveral other arrows of time have beenproposed, this thermodynamic arrowseems the least dependent on prior

assumptions [30, 31]. However, in the

context of our universe, it does de-

pend for directionality on recognition

of an initially constrained, dense,

high-temperature, high-entropy state,

whose evolution was triggered by the

Big Bang (Figure 6). With respect to

the subsequent expansion into free

space and the evolving universe as a

whole, this starting condition is con-

sidered to have been a low-entropy

state. The contrasting state is the

equilibrium condition of maximal en-

tropy, in which time has no signifi-

cance because all processes occur

with equal probability in either tem-

poral direction. Under equilibrium

conditions, the time symmetry of the

physical laws has its full expression,

but, ironically, loses all meaning. Any

perturbation of the equilibrium state

will introduce a Boltzmann asymme-

try and restart time. Paradoxically,

Boltzmann’s equation defining en-

tropy in terms of probability has no

temporal term. Its independence from

time is its claim to recognition as an

independent indicator. However, the

development of nonequilibrium ther-

modynamics [32, 33] introduced the

dissipation function, which is the rate

of production of “irreversible heat,”

the entropy production in excess of

the heat exchange associated with a

reversible process. The temporal state

of a system undergoing a reversible

process is similar to that of an equi-

librium state: the work potentially

available in one system is balanced by

that in another (formally in the sur-

roundings) through a coupling be-

tween them, and neither can budge.

Despite gradients that are available

for work, the system is metastable,

and no significant dissipation occurs

because change is infinitely slow. A

real change of state can occur only if

some of the potential work is dissi-

pated as heat lost to the surroundings.

The dissipation function provides a

measure of rate for a spontaneous

process. When applied to a system un-

dergoing a change in state, the dissi-

pation function gives the rate at which

FIGURE 6

The arrow of time. The arrow is shown as a spiral (represented by an -helix) passing from the pastthrough the “plane of present” to the future. The “plane of present” is represented by a spiral galaxy(with a period long compared to human timescales). Start- and end-points are the big bang, and“entropic death” (represented metaphorically by a detail from “Hell” by Hieronymus Bosch. Therewould of course be none of those discernable gradients in a universe at equilibrium). [Color figure canbe viewed in the online issue, which is available at www.interscience.wiley.com.]


the potential for work is wasted. If nocoupling occurs, all the potential work isdissipated as heat. Reversible transitionscan occur within a system in flux, be-tween states at similar potential, butthese will necessarily all be populated(equally if the work terms are zero, andflux coefficients high), and flux throughthe system requires a dissipative processat one end or the other.

Returning to the energetics of thebiosphere in this context, the major in-put of work comes from the light ab-sorbed in photosynthesis. Although theenergy is eventually dissipated as heat,thereby conserving the balance requiredby the first law, the strictures of the sec-ond law are temporarily smoothed out inthe more shallow dissipation functionprovided by conservation of energy in theweb of coupled electrochemical and met-abolic reactions. Indeed, the accumu-lated order and chemical complexity ofthe biosphere is a direct consequence ofthis entropic delay. Some of the inputenergy is borrowed for a while, and in-vested in life, in maintenance and evolu-tion of the biosphere.

THE ARGUMENT SO FARIt might be tempting to see in theapparent correlation between chro-nognostic range, and the increased in-formation content of the modern bio-sphere, some sort of reversal ofEddington’s arrow of time. The in-crease in information content in re-cent times appears to involve a highdegree of organization or negentropy,and, although we can’t talk about areversal of time, our perception oftime certainly seems to be extended ina fashion linked to this increase. How-ever, any further development of suchspeculation comes up against our lackof a philosophical apparatus for relat-ing semantic content to the thermo-dynamics of the physical world.

At this point, it is worthwhile to sum-marize the above discussion in the formof the following conjectures:

1. It is necessary to take seriously Shan-non’s distinction between two contri-butions to information processing: thesemantic content (or meaning) and

engineering problems inherent in in-formation storage or signal transmis-sion. Both aspects are essential. With-out the physical encoding, there wouldbe no message, without the messagethere would be no meaning.

2. Information storage involves changesin the physical medium for which thework could in principle be returned bycoupling to a suitable external system.These are well described by conven-tional thermodynamics. The integrityof information transmission can bescored by comparison between twosources using the probabilistic meth-ods of information theory.

3. Semantic content, in contrast, seemsto have no consequence that can bemeasured for the thermodynamiccontent of the system; it imposes noadditional thermodynamic burdenover that associated with the termsunder 1 and 2 above.

4. The semantic content has a value thatis entirely dependent on context. Thesignificance of the information onlybecomes apparent on translation.

These postulates apply to all formsof biological information from the DNAlevel to that of human intelligence.

In addition, we have noted the fol-lowing:

5. Translation of the semantic contentof the genotype leads to expressionof the phenotype. The emergentproperties of living things reflect thepossibilities for combinatorial com-plexity introduced by the transla-tional and interpretational steps. Anincrease in dimensionality is impor-tant in increasing combinatorial po-tential and hence complexity.

6. One characteristic of the phenotypeis a behavior that fits each species toan appropriate ecological niche. Thebehavior of living things implies asense of time or chronognosis. Chro-nognosis seems to be built into allliving things through biochemical tim-ers encoded by the genotype, but in-creasing complexity allows for moresophistication in temporal perception.

7. The ability to modulate behavior inthe light of successful prediction ofthe future has an obvious evolution-

ary advantage. There is a correlationbetween behavioral complexity, thecomplexity of sensory apparatus,and chronognostic range.

8. The evolution of H. sapiens sapiensintroduced a new development, thatof the human mind with the poten-tial for language, writing, etc., andthis has provided a pathway for se-mantic inheritance distinct from thatthrough DNA and opened new pos-sibilities for combinatorial complexityat the behavioral level. These in turnhave created opportunities for survivalin previously unexplored niches.

In order to appreciate the anthropo-centric perspective opened by that lastpostulate, we now need to review the de-velopment of our current views of time,civilization, and the place of man in thescheme of things. From the limitations ofspace and the author’s knowledge of his-tory and philosophy this will be brief andperhaps superficial.

CIVILIZATION AS AN EVOLUTIONARYPROCESS: HUMAN TIME AND SOCIALORDERThe origins of language are obscure; itseems likely that transmission of ideasbetween individuals and generationsdates back at least to the earliest tool-making cultures of 3–2 million yearsago. Starting with the development ofsophisticated languages and the migra-tions from Africa (Figure 7) around70,000 years ago,7 humans extended

7The migrations from Africa of the popula-tions from which the present distribution ofmankind is derived are illustrated by thedistributions of mitochondrial DNA types(for the female line), and Y-chromosomaltypes (for the male line), and the time scalederived from mutation rates (Figure 3).The migrations range over the period from�70,000 years before present (http://www.mitomap.org/WorldMigrations.pdf andhttps://www3.nationalgeographic.com/genographic/atlas.html). Modern languagegroups date from much more recent times[cf. [89]; and see articles in Science Vol.303(5662), Special Issue on Evolution ofLanguage, 2004].


their chronognostic range and biolog-ical competitiveness by creating newopportunities for combinatorial com-plexity at the social level. The basicbehavioral features involved, socialorganization in herd, pack, or tribe;communication between individualsthrough sound, scent, gesture, etc.;and alteration of the physical environ-ment to record status or range, arecertainly not unique to humans. How-ever, exploitation of the social oppor-tunities has depended on evolution ofseveral secondary features that, atleast locally, are unique to Homo s.sapiens. These include the anatomicalapparatus for sophisticated speechand a larger brain in which informa-tion about the environment can beabstracted and manipulated indepen-dently of the external world. Neithervocalization nor the mental functionsare unique. Use of sound in commu-

nication is common across the animalkingdom, and any dog owner, for ex-ample, will recognize that man’s bestfriend has an extensive repertoire ofassociative memories that provide agood description of aspects of theworld of importance to dog. However,in man, some attributes of the brainhave clearly been amplified. Records ofevents from the physical world arestored in an extended memory as co-herent “images” that retain some tem-poral and spatial context. These ab-stracted images of the outside worldcan reflect input through any one of thesenses and are partitioned in such away that parts can be compared andtheir associative properties modulated.Both the capacity and the informa-tional processing aspects have been ex-tended, introducing a far greater poten-tial for manipulation of the stored datain novel combinations. This allows de-

velopment over time of an internalmodel of the physical world that can beexploited for predictive purposes.

Some of the earliest evidence forabstract realization of the externalworld comes from the appearance offigurative carving and cave paintingssome 30,000 years ago (Figure 8). Theinternal world model is cumulative; asrecognized by Aristotle [34], the tablarasa of the newborn mind evolvesthrough a learning process. The co-herence of our conception of realitystarts in the womb and expands as ourphysical contact with the world makespossible an iterative process of verifi-cation and reformulation. In manyways, this learning process at the in-dividual level reflects the develop-ment of the human social condition.At the phenotypical level, ontogenyrecapitulates phylogeny; at the behav-ioral level, education recapitulates

FIGURE 7

The exodus from Africa. The origins of all modern human populations can be traced back to ancestors from Africa, through comparison of sequencesof mitochondrial DNA (female lines, or “Eve”, beige), and DNA from the “Y” chromosome (male lines, or “Adam”, cyan). Migrations from Africa dateback �70,000 years, providing an anthropocentric start to our history. The blue lettering indicates major mitochondrial lineages, the dark red italiclettering shows Y-chromosomal lineages, and the dark green numbers followed by kya show that arrival time in thousands of years ago. Only lineagesdating back to the period 35– 40 kya are shown), and the land masses and glaciation (white) reflect this time period. This Figure was constructed usingdata made available through the National Geographic Society’s Genographic Project web pages (https://www3.nationalgeographic.com/genographic/index.html), and this informative and interesting site is gratefully acknowledge.


civilization.8 In the development of

the individual, as in the advancement

of civilization, the acquisition of lan-

guage is paramount. With language,

the advantages of group living are ex-

panded in a multiplicative fashion.

Sharing of ideas allows sharing and

accumulation of skills, teamwork in

the hunt and in defense or attack, and

the development of social skills, social

cohesion, schooling, and oral history.

THE DEVELOPMENT OF COMPLEXSOCIETIESAlthough its origins are hotly debated, the

development of agriculture—the planting

of crops, and animal husbandry—is

thought to be more recent, though it was

well established 10,000 years ago [35]. Ag-

riculture led to more organized societies,

division of labor, property, trade, etc. The

development of social order encouraged

the extension of counting to accounting

and development of a number system

and records. Number systems in turn al-

lowed dissociation of the concept of

numbers from specific association with

things, and hence their more abstract use

in calculation. Recording led to writing,

pictograms to ideograms and phono-

grams, and a more abstract representa-

tion. Parallel developments included rec-

ognition of the importance of prediction

of seasonal effects to agriculture, which led

to astronomy-based calendars. The devel-

opment of archival records and the inven-

tion of predictive devices on a large scale

(monumental altars, pyramids, etc.) were

associated with a stratified social order with

castes, including that of the scholar-priest-

scribe. With language, the abstraction ofideas through writing and arithmetic andtransmission of ideas between individualsand over generations, the process of learn-ing was extended beyond the individual tothe social and beyond the individual lifes-pan to the duration of society. It led even-tually to sophisticated societies.

The point I want to bring out here isthat these developments were incre-mental. Both the phenotypical changesleading to the development of modernman as a species and the behavioralchanges leading to modern society wereall part of evolution and involved manysmall changes occurring in parallel toprovide a continuity of development, allof which has been in the context of abiosphere also evolving in parallel.

THE SUPRA-PHENOTYPICALAlthough the development of civilizedsocieties was an evolutionary process,the developments leading to such soci-eties are different in kind from the bio-logical features generally lumped to-gether as the phenotype. Withoutintending to imply anything unnaturalabout these extensions of biologicalfunction, it might be useful to recognizethe difference as supra-phenotypical(Figure 9), meaning, minimally, that itinvolves the conscription of the inani-mate world to the service of biologicalbehavior. We can recognize that exten-sions of behavior that make use of in-animate matter, Dawkins’ extendedphenotype [36], are common in the an-imal world— use of tools, decoration,nest-building, beaver dams, etc.—andcould even be extended to the plantworld, where plant form takes everypossible advantage of geology or localstructure. However, it is in extension ofthe supra-phenotype that our specieshas made its mark. In the case of hu-man behavior, this has involved sym-bolic representation of mental images,abstracted from their biological contextin the brain (pictorial images, language,counting, etc.) and recorded in external,inanimate form (art, writing, numbers,etc.). Development of abstract repre-sentation and the parallel developmentof abstraction at the level of mental pro-cessing seems to involve a self-amplify-

8Ernst Haekel’s famous phrase “ontog-eny recapitulates phylogeny” was an im-portant topic of discussion in late 19thcentury biology, and, although the de-tails of Haeckel’s embryology were wrongand the pedants have scoffed at earlierinterpretations of this idea, it continuesto be fruitful (cf. [64]). The notion thateducation recapitulates civilization, inthe context of development of the mind,was central to Freudian psychology, andhas been further developed by other psy-chologists and in Piaget’s writings aboutchildhood development.

FIGURE 8

Lascaux cave drawings, and similar cave art and figurines from the pre-historic era, represent earlyexamples of the representation in physical form of semantic content abstracted from the mind (imagefrom Wikipedia, http://en.wikipedia.org/wiki/Image:Lascaux2.jpg) [Color figure can be viewed in theonline issue, which is available at www.interscience.wiley.com]


ing loop, through which the sophistica-tion in processing of a cumulative worldmodel is exponentially increased. Am-plification occurs at both the individualand societal levels. Through use of theseabstractions, we have extended boththe limits of our experience and ourchronognostic range. On the morepractical side, our manipulation of theinanimate world has given us engineer-ing, architecture, weapons of war, etc.Perhaps the most momentous conse-quence, as noted briefly above, is thatwe have introduced new forms of inher-ited semantic content that are unique(at least locally) to our species. This se-mantic inheritance at the behaviorallevel is quite distinct from the semanticcontent carried by the genotype.

The ordering of society made possi-ble by the inventions briefly outlinedabove was a necessary part of the devel-opment of social structures underwhich these inventions could be ex-tended and exploited. It is customary toview this process as one of establishingstability. However, organized socialstructures are necessarily highly or-dered and hence inherently unstable.Their fragility is all too apparent fromhistory and especially in this age of ter-rorism. The appearance of stabilitycomes from the development of politi-cal, social, economic, technical, andphilosophical tools that make possiblethe continued maintenance, renewal,and evolution of these institutionsthrough the input of work.

COUNTING TIMEThe “message” encoded in DNA hasvalue only in the context of life and ofbehavior. Similarly, our daily conversa-tions have a value only in the context ofan evolved intelligence and civilization. Adetailed examination of the extensions ofbehavioral complexity leading to ourmodern world is beyond the scope of thisessay. However, the major part of theprocess has been in the semantic realmand has been linked to our developmentof a sophisticated appreciation of time.

FIGURE 9

Scheme to illustrate the levels of combinatorial complexity arising from increased dimensionality following semantic translation at the somatic level (top). Theresidual information content refers to that originating from the genome at conception; in metazoans, the mature phenotype obviously contains many diploidcopies, but most of these are not involved in reproduction, and only half the original is passed on in each haploid gamete. The estimated number of proteincoding genes (�24,000) represents �2.5% of the genome. A smaller fraction encodes RNA directly involved in the translational machinery. A much largerfraction (probably �80%) is transcribed, and from its characteristics is likely involved in determination of tissue specificity, ontogenic sequencing, etc. For arecent review, see 104. The combinatorial possibilities expand with dimensionality more dramatically than shown; the vertical axis of the schematic should bethought of as on a log scale. Complexity inherent in language (bottom). The representation of complex thought in literature, conversation, etc. depends on thecombinatorial exploitation of a relatively small number of symbols in sentences whose structure encodes meaning. The ideas (memes) are the units of semantictransmission, and the complexity of our culture depends on the translational processing and mutation in individual minds (not shown). Similar schemes couldbe shown for mathematics, logic, music, etc. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

The “message” encoded in DNAhas value only in the context of

life and of behavior.


The evolution of our ideas about man’stemporal place in the scheme of thingshas seen an interesting interplay betweensocial order, religion, and scholarship.9

Our recent chronognostic history islinked to the history of numbers. Perhapsthe earliest records indicating formal rec-ognition of extended time date back�20,000 years and are represented bymarkings on sticks and bones from Eu-rope that are thought to record the daysbetween successive new moons. The ear-liest Egyptian calendars date back to�4500 BC, and by 2700 BC they had es-

tablished a year of 365 days. We owe ourdaily time scale of 24 hours, divided into60 minutes of 60 seconds to the Sumer-ian-Babylonian sexagesimal number sys-tem from the same period. Geometryevolved in the context of engineering. Thecuneiform tablets from Sumeria andBabylon (1900–1600 BC) are mostly eco-nomic records, but some also show aknowledge of Pythagoras’ theorem thatpredated the Greek discovery (�500 BC)by at least several centuries (Figure 10).They also demonstrate the solution ofquadratic (and higher-order) equations,and their application to land manage-ment. The Ahmes and Moscow papyrifrom �1850 BC include examples of cal-culation of areas of circles and volumes ofpyramids. The construction of the pyra-mids, some of them a thousand years ear-lier, must have required a fairly accurateknowledge of the properties of triangles.Although less sophisticated, Stonehenge(phase I, �2950 BC), with its astronomi-cal orientation, dates from about thistime. Similar mathematical develop-ments occurred independently in Chi-nese and Mayan cultures. The earliest In-dian written texts, the Vedas, date from

15–5th century BC, and contain the Sul-basutras from �800 BC as mathematicalappendices. These include details of thecalculations needed for correctly orient-ing altars, including Pythagoras’ theorem.The Indian decimal number system fromthis era evolved, with place dependence,zero and all, to give us, via the Arabs, ourpresent decimal number system.These mathematical tools became im-portant in the development of astro-nomically based calendars, which re-quired records kept over extendedperiods, and precise geometric mea-surement. Because of the importance toagriculture, the application of theseskills to engineering of irrigation sys-tems and the prediction of seasonsbrought to those able to make suchmeasurement both power and prestige.The development of religious ideaslikely evolved with language and socialstructure,9 but no doubt the expandingtemporal perception arising from thedevelopment of calendars, and the ap-preciation of power series, openeddeeper speculation about the underly-ing mysteries. What is of interest here isthe emergence of the scholar-priest-scribe caste in which the tools of writ-ing, reading, and arithmetic that facili-tate the development of abstract ideasand a perception of time, were appliedin both a practical and a religious back-ground. Many of the records that havesurvived seem also to have served ateaching function, indicating that for-mal institutions for transmission ofknowledge evolved at the same time.

THE “LIMITS OF EXPERIENCE” ANDTHE BIRTH OF MODERN SCIENCEThe development of abstract number sys-tems, and the ability to manipulate scaleby power series, let to the idea of infinityand recognition of the philosophical dif-ficulties introduced by this. Parallel devel-opments in different philosophical sys-tems led in different directions. In theframe of Western civilization, startingperhaps with Parmenides, the Greeks de-veloped a dialectic approach that sepa-rated the philosophical problems frommythology, and effectively partitionedspeculation into temporal and eternal do-mains appropriate, respectively, to man

9The common roots of religion, philoso-phy, and science in social behavior wasintroduced by Durkheim (see http://durkheim.itgo.com/religion.html), whois considered by many as the father ofsociology. He suggested that these all de-rive from the need for groups to find acommon social identity through sharedparticipation in ritual, creation myths,and exploration of man’s connectednessto nature. Robert Alun Jones [90] has anice critique of Durkheim’s The Elemen-tary Forms of Religious Life [91].

FIGURE 10

Cuneiform tablet from Ancient Babylon (�1700 BC) showing Pythagorean triplets. Plimpton 322 fromthe Plimpton collection (Columbia University) (image from Wikipedia, http://en.wikipedia.org/wiki/Image:Plimpton_322.jpg). For much of the information about the development of mathematics, I amindebted to the MacTutor Web site (see Supplementary Sources for URL).


and the Gods. In this divide, man’s placewas constrained by the limits of experi-ence. The ideas of Plato and Aristotle, asmolded by Plotinus, were co-opted by theearly Christian church, and the difficultyof placing Jesus in both temporal andeternal realms lead to an evolution fromneo-Platonic views to the concept of theTrinity. Despite an extraordinarily liberalpronouncement in the Edict of Milan,10

the emperor Constantine later proclaimedthe Christian faith as the state religion. Thehot debate within the Church about thenature of the Trinity, which dominated thetheology of this period, was formally set-tled at the Council of Nicaea11 in 325 AD(Figure 11), and its aftermath,12 leading tothe Council of Constantinople in 381.These led to the triumph of the homoou-sian view and established orthodox Chris-tianity, but also led to suppression ofother beliefs as heresies. Perhaps inevita-bly, the power brought to the Church by itscentral politic status eventually placed it inthe position of protector of the status quo.A hundred years after the Council of Con-stantinople, following invasions byFranks, Goths, Saxons, Huns, and Van-dals, the Western Empire was in ruins.The Church kept the light of Western

scholarship alive through the Dark Ages,

but the dual role as state religion and

repository of knowledge persisted and led

ineluctably to defense of the dogma,

which largely froze philosophical devel-

opment in the Western world for over a

thousand years. The development of mod-

ern science from Roger Bacon, through Co-

pernicus, Bruno, Galileo, Newton, Spinoza,

Darwin, and Einstein, to name but a few,13

FIGURE 11

The Council of Nicaea. Bishop Nicholas forcefully argues for the homoousian cause (Sistine Chapelfresco). The transformation of St. Nicholas into Santa Claus is a nice example of memetic mutation (seetext). Image courtesy of St. Nicholas Center (www.stnicholascenter.org), which has a nice account. Theforcefulness of Bishop Nicholas was not just verbal! [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.]

10Constantine was Emperor of the West.As noted by Gibbon (1737–1794): “Thetwo emperors proclaim to the world thatthey have granted a free and absolutepower to the Christians, and to all others,of following the religion which each indi-vidual thinks proper to prefer, to which hehas addicted his mind, and which he maydeem the best adapted to his own use.They carefully explain every ambiguousword, remove every exception, and exactfrom the governors of the provinces a strictobedience to the true and simple meaningof an edict which was designed to establishand secure, without any limitation, theclaims of religious liberty.”11Again from Gibbon: “The Greek wordwhich was chosen . . . bears so close an af-finity to the orthodox symbol, that the pro-fane of every age have derided the furiouscontests which the difference of a singlediphthong excited between the Homoou-sians and the Homoiousians.” The former,meaning in this context that the Father,

Son, and Holy Ghost were of the samesubstance (consubstantial, co-eternal),was the essential description of the natureof the Trinity that became the orthodoxfaith; the latter, meaning of similar sub-stance, was preferred by the Arian faction.Supposedly, the distinction was lost on themajority of Bishops at the Council of Ni-caea, hence ‘not an iota of difference’.12The Nicene Creed was meant to settlethe issue, but the argument was far fromover. Constantius, who followed Con-stantine as Emperor, supported the ho-moiousian view. It was hotly contestedby Athanasius, Archbishop of Alexander,but with the reign of Julian and the re-turn of paganism, he was exiled (GoreVidal gives a readable account of thisperiod in his historical novel Julian). Hewas restored to the archbishopric by Jo-vian, and his views waxed and thenwaned again under Valens. The ortho-dox homoousian view, summarized inthe Athanasian Creed, eventually pre-

vailed under Theodosius, with the edictsof the Council of Constantinople. Thedebate has continued, for example, thehomoiousian view was revived, under Jef-ferson’s influence, in the Unitarian splitfrom Puritanism in the United States.13Each of the individuals named hereencountered resistance from establishedreligion for their temerity in proposing toshift the balance. In all cases, the au-thorities could have followed a less dog-matic course. The philosophical distinc-tion deriving from the neo-Platonists,which separates the realm of religionfrom the “limits of experience” of man,was well recognized in the early Churchand could have been invoked to ease thetransition from dogma. For example,Pope Clement IV rescued Roger Bacon(1214 –1292) from obsurity and incar-ceration, but on Clement’s death, Baconwas imprisoned again for much of therest of his life, charged with suspectednovelties in his teaching. (Bacon, and


can be seen as a wresting from theChurch of its claim to the farther reachesof time and space, through extension ofthe limits of experience of man.

The perception of time as proceed-ing from the past, through the presentto the future, is neatly summarized bythe aphorism, “We can know the pastbut we cannot change it; we can changethe future, but we cannot know it.” Wemay reasonably assume that temporalpolarity applies precisely in the physicalworld, based on time-invariant physicallaws, the Big Bang, and Eddington’s ar-row. But our subjective interpretation oftime in the world of experience is obvi-ously conditional; our present view isbased on quite recent philosophic de-velopments. The “past” changed dra-matically with the introduction of evo-lutionary time. Starting with Lyell’swork on geology and the work on evo-lution from Darwin and Wallace, ourview backward has changed from Arch-bishop Ussher’s creation of the world in4004 BC, to that of the Big Bang, some12–14 billion years earlier. As for thefuture, prediction of cyclic phenomenaon an astronomical scale developedquite early; the astronomers of the earlyBabylonian (Thales likely had somehelp from this quarter) and Chinese civi-lizations, and independently, the Maya,were able to predict solar eclipses. New-tonian mechanics extended the range tocomets and the orbits of Neptune and

Pluto. From Einstein and the revolutionin quantum mechanics leading to nu-clear chemistry, we can now under-stand the life and death of stars andgalaxies. Everyday life is dominated bymore chaotic notions, but these arelargely smoothed out by the highly or-dered societies in which we participate,so that we can anticipate a trajectory ofour lives (‘barring accidents’) with someconfidence. As I sit at my desk in theattic, I can watch the last 16 frames ofthe radar coverage on my computer andreckon with complacency that the tor-nado cells in the storm front will pass 20miles to the south.

To return to my main theme, it willbe apparent from the above that therelation between information content,time, and energy is quite different fromthat of our earlier speculation in rela-tion to the arrow of time. The semanticcomponent of information content hasno conventional entropic implicationsand hence no direct relation to thermo-dynamic energy and the physical world.On the other hand, the depth of ourperception of time has everything to dowith the extension of behavior into thesupra-phenotypical range; the develop-ment of language, number systems andabstract thinking that accompaniedevolution of the brain’s capabilities; andthe accumulation of learning, and theseare the features that have enabled us todirect energy into the molding of our

societies. The present adult mental abil-ities of our species are predicated onthe combined efforts of thousands ofprior generations of our kind in estab-lishing the external semantic heritagethat provides the base and frameworkof our knowledge. That cultural devel-opment, just as our education, has nec-essarily been sequential and cumula-tive. However, the phenotype hasprobably not changed significantly overthis period: the youth who sets off tocollege today, armed with the rudi-ments of calculus and modern biology,a smattering of some foreign language,and the beginnings of an appreciationof literature, art and music, is the sameyouth who knapped flints for his spearor who scraped her skins at the firewhile listening to the wisdom of theshaman 30,000 years ago. The mind thatshaped the art of Lascaux was not infe-rior, but educated to a different context.

SOME PHILOSOPHICAL IMPLICATIONSFor a neophyte to philosophy, com-pelled by this question of the thermo-dynamic status of semantic content torevisit the rudimentary introduction de-rived from undergraduate days, it hasbeen salutary to find that the themesintroduced in the preceding sectionsrun through philosophical thoughtfrom the earliest days. A seasoned phi-losopher might point out that my dis-cussion so far represents a banal exer-

Grosseteste, his mentor, were the fathersof the scientific method; http://www.fordham.edu/halsall/source/bacon2.html).

Galileo, in his letter of 1615 to the GrandDuchess Christina of Tuscany (Christinede Lorraine; http://www.fordham.edu/halsall/mod/galileo-tuscany.html) artic-ulates a straightforward realist view:“But Nature, on the other hand, is inex-orable and immutable; she never trans-gresses the laws imposed upon her, orcares a whit whether her abstruse rea-sons and methods of operation are un-derstandable to men. For that reason itappears that nothing physical whichsense-experience sets before our eyes, orwhich necessary demonstrations prove

to us, ought to be called in question . . .,”and invokes with great eloquence theChurch Fathers and St. Augustine as rec-ognizing the need to reconcile the teach-ings of the church with natural experi-ence. It was to be a long time before thisview prevailed. The current Catholicperspective is summarized in the addressof Pope John Paul II to the PontificalAcademy of Sciences on the occasion oftheir report on the “Galileo case” (1992).This affirms St. Augustine’s view: “If ithappens that the authority of SacredScripture is set in opposition to clear andcertain reasoning, this must mean thatthe person who interprets Scripture doesnot understand it correctly . . ..” The ad-dress also contains a somewhat belated

acknowledgement of the philosophicalcorrectness of Galileo’s position: “Paradox-ically, Galileo, a sincere believer, showedhimself to be more perceptive in this re-gard than the theologians who opposedhim . . .. We also know of his letter toChristine de Lorraine (1615) which is likea short treatise on biblical hermeneutics.”(An English translation of the address isavailable at http://www.its.caltech.edu/�nmcenter/sci-cp/sci-9211.html.)

The distinction between religion andthe “limits of experience” has been re-formulated in a contemporary setting inJudge Jones’ ruling on the Intelligent De-sign case (Kitzmiller v. Dover School Dis-trict; http://www.pamd.uscourts.gov/kitz-miller/kitzmiller_342.pdf).


cise: the regurgitation of commonplaceideas from a mishmash of science, phi-losophy and history. But isn’t that howit has to be: a viewpoint that arises nat-urally from my argument? Because ofthe role of translational processing ininterpretation of semantic content, themind can only accommodate ideas thatit is already equipped to deal with. Theadage that there are no new ideas, just arehashing of the old, becomes a truism.However, it is in the rehashing thatprogress is made, and Shannon’s dis-tinction between the engineering prob-lems and semantic content provides afresh perspective on this process. From

the arguments above, it should be clearthat any “idea” entering our conscious-ness does so as a result of translationalprocessing, and our appreciation de-pends on a linkage to other ideas al-ready in residence. The availability in aparticular mind of a particular combi-nation of observations and old ideas ina particular spatiotemporal context isunique to a particular individual, andthe re-ordering of those ideas in newpatterns depends on the particular ex-perience of that individual and on crit-ical choice. It is this re-ordering, the“mutation” of old ideas, that is a poten-tial source of philosophical renewal,

providing the grist for the evolutionarymill. Whether or not the new combina-tions are useful or have any originalmeaning is determined by their reso-nance in a wider social context.

EVOLUTIONARY EPISTEMOLOGYThe philosophical perspective devel-oped here has clear parallels in Pop-per’s Three Worlds model and his ob-jective knowledge epistemology [37,38], Campbell’s evolutionary epistemol-ogy [39], and Dawkins’ memes [13, 36,40].14 My supra-phenotypical semanticinheritance is recognizable in Popper’sextrasomatic World 3 (Figure 12), butarrived at from a new direction. How-ever, the “insubstantial” nature of se-mantic content makes possible a clearerdistinction between the mind, and boththe phenomenal world and that of theexternal semantic inheritance. The se-mantic traffic between the three worldsreflects a sophisticated translation pro-cess that is not well understood. How-ever, as already recognized by Aristotle,the higher processing through which weinterpret primary sensory input seemssimilar to the processing through whichwe interpret “images” stored in thebrain, or the semantic content of theexternal semantic heritage, in books,movies, art, science, math, music, po-etry, etc. The data from the real worldthat impinge on our sense organs arepurely physical, though they can act ascarriers of messages. What we describeas entering our minds, what appears tous as perception, and what becomesmeaningful is the result of translationalprocessing at many levels, from the in-nate on up.15

Similarly, because of the insubstan-tial nature of semantic content, all per-

14In summary, memes are self-replicat-ing mental agents: “The haven allmemes depend on reaching is the hu-man mind, but a human mind is itselfan artifact created when memes restruc-ture a human brain in order to make ita better habitat for memes. The avenuesfor entry and departure are modified tosuit local conditions, and strengthenedby various artificial devices that enhance

fidelity and prolixity of replication: na-tive Chinese minds differ dramaticallyfrom native Frech minds, and literateminds differ from illiterate minds. Whatmemes provide in return to the organ-isms in which they reside is an incalcu-labe store of advantages—with someTrojan horses thrown in for good mea-sure. . .” Daniel Dennett [56] (quoted byDawkins in Viruses of the Mind [40].

15In addition to Dennet’s treatment [56],

further complexities of the human brain

with its many levels of processing are

discussed in the context of the contribu-

tions of emotions to homeostasis and the

wider social milieu by Antonio Damasio

[92] in Looking for Spinoza, which also

has a nice account of Spinoza’s life.

FIGURE 12

Popper’s Worlds 1, 2, and 3, expanded into three dimensions. We see ourselves as individual minds(World 2). Our perception of the physical World 1 might be unique (arrows from left showing sensoryinput), but we all experience the same world, tied by physical laws. Our “conversations” (centralcolumn) involve exchanges in which translational processing, access to associative memory, mutationof ideas, etc. are so rapid as to appear automatic. The semantic heritage of World 3 (right) appearsdynamic, but is static except for the horizontal exchanges with individual minds. The Internet traffic isa relatively new feature, but we depend on a copying function that is error-free, so the verticalexchanges indicated do not change the heritage. However, the speed of exchanges via the Internet iscontributing to an acceleration in exploration of combinatorial possibilities made possible by thehorizontal exchanges with minds. [Color figure can be viewed in the online issue, which is availableat www.interscience.wiley.com.]


ception of the external semantic heri-tage must also be the result oftranslational processing of data intro-duced via a thermodynamic vehicle.The semantic heritage itself is passive inthe sense that all the apparently dy-namic aspects of World 3 are engen-dered by translational traffic betweenindividual human minds and betweenminds and the knowledge banks. De-spite the enormous impact of the se-mantic inheritance, the human mind isstill the only vehicle for active change inthe semantic heritage.16 The factorsthat guarantee further development arethe uniqueness of each mind, access,and selection.

The information storage and re-trieval aspects of the “engineering side”pop up in different disguises: as DNAprocessing in World 1, as memory andprocessing in World 2, and as the manyinvented methods for storing informa-tion that stabilize the semantic heritageof World 3. However, the semantic con-tent in the brain must be as insubstan-tial as semantic content in DNA, or onpaper or in digital format. It is availablefor examination only because it can betransiently trapped in a thermodynamicform. The distinguishing features on theengineering side are the mechanismsfor handling storage and transmission.As Popper, Campbell, and Dawkinshave all emphasized, both at the DNAlevel, and in the brain, two aspects ofinformation transfer are essential for

evolution to occur: a high degree of fi-delity in archival storage and transmis-sion and a degree of mutability to pro-vide the working stuff for selection.From the wider evolutionary perspec-tive, life can be seen as a set of mecha-nisms, driven mainly by absorption oflight in photosynthesis, through whichthermodynamic potential can be ex-ploited for self-replication, throughtranslation and transmission of seman-tic content, with selection of spontane-ous variants to provide the mechanismfor change. For most of the biosphere,the semantic content involves DNA,and its direct or indirect translationalproducts. For human kind, the wholeapparatus of culture and civilizationprovides an additional channel. Our“advance” over the last few millenniahas come about because of evolution inthis latter mode: the supra-phenotypi-cal semantic heritage.

PARALLELS WITH POPPER’SOBJECTIVE KNOWLEDGEArguments about the nature of realityhave occupied a central position inphilosophical debate at least since pre-Platonic days [34, 37–39, 41, 42]. Itseems to me quite interesting to revisitthis enduring argument in the light ofthe dual aspects of information content.The struggles of Bishop Berkeley withthe world as illusion, and Dr. Johnson’scry of pain as he kicked the stone,17

seem freshly illuminated by the light of

Shannon’s distinction between seman-tic content and the engineering prob-lem. The evolutionary perspective de-veloped above lands me firmly in therealist camp as expounded by Popper/Campbell/Bartley (though other philos-ophers seem to view realism throughseveral different prisms). We might sup-pose that, in the light of the insubstan-tial nature of semantic content, our in-dividual zeitgeist is a transparency. But,of course, this is only part of the picture.The mental constructs through whichwe recognize the world are the productsof a mind that owes both its being andbecoming to the biological context of anevolved phenotype and the socio-eco-logical context of a civilization. In thiswider framework of the shaping of thehuman mind, we can perhaps, withoutinsult to Descartes, extend his idea toembrace our wider debt: cogito ergo su-mus brings the rest of us into the pic-ture. Necessarily, our debt extends backthrough time and to the whole biosphereand embraces the physical world.

The lack of a thermodynamic con-text for semantic content raises episte-mological questions, both in relation toour perception of everyday reality, and

16We should note that this unique posi-tion may already be shifting as comput-ers take on a more direct role in inter-pretation of the physical world, throughdeposition of data in the knowledgebases, and through predictive calcula-tions, in molecular dynamics modelling,for example. The question of whethercomputers can be designed to emulatethe human brain has been much dis-cussed. Both Penrose [69] and Putnam[34] reject the idea and certainly make astrong case against believing that a clas-sical Turing machine can perform thisfunction. In so far as a Turing machineis designed to be error-free, I wouldagree. On the other hand, it seems that if

a computer is to emulate the mind, itshould be treated like a newborn childand allowed to learn about the world,make mistakes, feel pain and pleasure,and develop its own opinions. Withoutthe evolutionary potential for “error”—some capacity for the mutation of se-mantic input—a computer will never beable to contribute original ideas. Ofcourse, these requirements provide somehardware and software challenges, butsteps in this direction are underway [93].17From Boswell’s Life of Johnson: “After wecame out of the church, we stood talkingfor some time together of Bishop Berkeley’singenious sophistry to prove the nonexist-ence of matter, and that every thing in the

universe is merely ideal. I observed, that

though we are satisfied his doctrine is not

true, it is impossible to refute it. I never

shall forget the alacrity with which John-

son answered, striking his foot with

mighty force against a large stone, till he

rebounded from it—‘I refute it thus.’ ” The

recent series of movies starting with “The

Matrix” provides an exploration of “Brain

in a Vat” themes that follow an anti-real-

ist approach similar to Berkeley’s, but with

a machine intelligence replacing God (see

Hilary Putnam [34] and http://www.

uwichill.edu.bb/bnccde/ph29a/putnam.

html for the brain-in-a-vat reference).

The lack of a thermodynamiccontext for semantic content

raises epistemologicalquestions . . .


in relation to the nature of the scientificenterprise. Although all hypotheses aregenerated in the semantic realm, theyare tested in the observable, phenome-nal, thermodynamic world, which pro-vides the context for their evaluation. Ahypothesis that fails to account for well-authenticated properties of the observ-able world needs to be discarded ormodified. Popper developed this ideathrough his falsification approach. Pop-per’s view was largely a response towhat he saw as a crisis of “certainty” inphilosophy, which arose from assaultsfrom many fronts on the deterministicview of science and its philosophicalunderpinnings following the successesof Newtonian mechanics. Demonstra-tions of the limits of inductive logic dat-ing back to Hume (or earlier, to theskeptics), the introduction of Darwin’sideas, the beginnings of psychology, thelimitations of language, and the failureof logical positivism to find a solution tothe problem of “truth,” were com-pounded by the difficulties introducedby quantum uncertainties and relativis-tic views of space and time. By the mid-20th century, a philosophical view wasestablished (associated with contribu-tions from, among others, Frege, Car-nap, Quine, Godel, and the early Witt-genstein [43]), that verification was notan appropriate target for the scientificenterprise. Popper [44] pointed out thateven though this may be so, we could atleast know when a hypothesis was in-adequate, because of its failure in pre-diction of future observation. Heframed this idea through his falsifica-tion criterion: a hypothesis has valueonly if it does account for “the facts,”and if criteria for testing are built in as“risky” predictions. Popper recognizedthe need for some epistemologicalfoundation and suggested that thiscould be based on the recognition of anexternal observable world governed byinvariant laws: his World 1. He found

support for such a base in biological

evolution and in the extension of evo-

lutionary ideas to the domain of hy-

pothesis testing. Popper believed that

science could establish a reliable de-

scription of this world by an iterative

process of investigation based on a test-

ing against observable reality. He got

into trouble with his philosophical col-

leagues [34, 45, 46] on a number of

counts, to some extent because his crit-

ics failed to notice his distinction be-

tween the logic of the falsification ap-

proach and the practice of working

scientists. Popper recognized that the

latter was inevitably more complex be-

cause of practical difficulties and fac-

tors such as human fallibility and the

imprecision of the knowledge base. But

to the pedantic, examination of histori-

cal cases suggested that the mechanism

of falsification just didn’t always work

as advertised; in practice, a strong hy-

pothesis could survive after falsification

because it still seemed to work pretty

well. At a more formal level, following

Tarski’s formulation of a semantic the-

ory of truth [47], Popper attempted to

extend the idea of falsification by sug-

gesting an algorithm for scoring the

truth content of a hypothesis, which

was suggested to contain inconsisten-

cies [34, 45, 46]. Another area in which

Popper got onto trouble was in use of

his falsification approach to provide a

demarcation between science and non-

science. Popper proposed that a good

scientific hypothesis must include pre-

dictions that allow unambiguous tests.

If a hypothesis was framed in such a

way as to allow wriggle room through

ad hoc additions that could explain

away inconsistencies: it wasn’t science.

His critics pointed out that science in

practice often adopts the latter strategy

as an expedient for progress. An alter-

native line of demarcation can be

framed in the light of the second prin-

ciple suggested in my statement of pur-pose above.

“RELATIVE TRUTH”An alternative scenario for revolutionsin science was introduced by Kuhn [48],in which normative science—successfulscience, conducted according to estab-lished paradigms—accounts for themain flow of the scientific enterprise,but gets shocked out of its rut by para-digm shifts that are incommensurablewith current thinking and which arerare and disruptive. A period of disputefollows after which social consent de-termines acceptance (or otherwise) ofthe shift, and the system settles backinto a productive complacency. In myview, both Popper and Kuhn havesomething interesting to contribute.Kuhn recognized the importance of asocial element, and his revolutionaryscenario certainly fits some major shiftsin scientific opinion, including Mitch-ell’s chemiosmotic revolution. Theproperties of semantic transmission,the nature of our daily “conversations,”and the mechanism of education dis-cussed above, lead to the necessity of asimilar social involvement. However,Kuhn scants the epistemological com-plexities, which Popper deals with quiteeffectively. The philosophical communityseems to have decided that the two viewswere contradictory rather than comple-mentary, projected this as a Kuhn/Popper war, and declared Kuhn thevictor. The outcome seems to havebeen, at least in the United States, thatthe mainstream of philosophical dis-cussion has passed Popper by, thebaby of falsification, and the evolu-tionary framework, have both beenthrown out with the bathwater, andhis views have been eclipsed (but seeMiller [49] for a defense).18

An influential scenario favored insome circles has been a view of truththat removes it from any firm pinning to

18See for instance the books by Kuhn[48], Rorty [52], Kenny [43], Kripke [94],Feyerabend [46], and Bloor [50]. Putnam[34] has an equitable critique of Rorty.For a more direct response to Bloor, see

Slezak [95] (also available on-line at:http://hps.arts.unsw.edu.au/hps_content/staff_homepages/p_slezak_site/Article%20Links/1994/Slezak_Bloor_2nd_Look_1994.pdf). The antirealist, constructivist,

view has also been championed by Fine[51], with more thought, but still throughthe prism of a highly restricted correspon-dence theory of truth.

In my view, because of the problems in


external reality [46, 50, 51]. Kuhn’s so-cial context has been combined with amishmash of ideas from the later Witt-genstein to produce a distortion of thescientific method, most notably in theself-serving “Strong Programme in theSociology of Knowledge” from Bloor[50] and the Edinburgh group, in whichcongruity with reality is abandoned andtruth is relative and determined by so-cial consent. At least some blame mustalso attach to Rorty [52], who took asimilar neo-pragmatic approach thatrejected realistic philosophies. He re-gards the scientific enterprise as agame, useful perhaps, but beyond thedignity of the serious philosopher. Rortydefines realistic philosophy in terms ofa correspondence theory of truth, butthis seems a bit of an Aunt Sally, whoseknock-down has been used to tar allrealistic philosophies. Feyerabend’s se-mantic instability [46] recognized themutability of ideas, but he later ex-tended his thinking in an anarchisticdirection that also contributed to thefashion for “relative truth.” Popper hasbeen viewed as adhering to a corre-

spondence theory of truth, but his atti-tude was more provisional. His em-brace of falsifiability was explicitly inresponse to recognition of the slipperynature of truth—the limitations of in-ductive logic preempt the view requiredby a strict correspondence theory—andTarski’s resort to meta-languages wasitself an attempt to deal with these dif-ficulties. Popper’s evolutionary episte-mology seemed to provide a pathwayout of this impasse. Meanwhile, theogre of relative truth has permeatedmuch of our culture and social thinking,and has, in the Bush administration,been incorporated quite explicitly intopolicy at the highest political levels [53],leading to a noticeable decline in thenation’s standing among the interna-tional community.

THE IMPACT ON SCIENCEMost scientists are unaware of this dis-jointed view of truth, seen as a “philo-sophical horror show” by Putnam [34],and soldier bravely on in Popper’sblithe spirit, mainly because it reflectswell their everyday experience of the

utility of a testing of hypothesis. In myview, this uncomplicated approach isquite appropriate; formalization of ‘thescientific method’ is difficult because“scientific knowledge” is a moving tar-get. The practice of science occurs in acontext of observation, hypothesis, andmeasurement, in which incompleteknowledge is the norm. Technical ad-vances allow the application of newmethods of measurement, which intro-duce a new observational backgroundagainst which hypothesis must betested and renewed, refined or dis-carded. All this is conducted in an en-vironment of peer review which can befar from ideally disinterested. The ac-count by Gilbert and Mulkay [54] ofthe controversy leading to eventualacceptance of Mitchell’s chemiosmotichypothesis provides an interesting over-view of the sociological side. Neverthe-less, the method of science works in thelong run; viewed in the evolutionarycontext, it is clear that science has aidedthose societies that have recognized thebenefits in exploitation of resources;economic prosperity and survival through

philosophy identified by the mid-20thcentury, discussions of the nature of truthin the philosophical, scientific, or everydaymeaning, are bound to be somewhat fu-tile, even when conducted against an evo-lutionary background. To this extent, apedantic adherence to a “correspondencetheory of truth” is indefensible. However,this does not mean that we have to aban-don attempts to relate experience to anexternal reality. To avoid the pitfalls ofinductive circularity, an external realitygoverned by invariant laws should per-haps be presented formally as a hypothe-sis, in which our understanding is provi-sional. Since the assumption of anexternal reality forms the external peggingfor most of the inductive circles that formour knowledge base, I believe it is fairlysecure. A statement that “T is true” thenrepresents a hypothesis that T is congruentwith that reality. Perhaps we can pin dis-cussion down by acknowledging a realitythat allows for an observation-based ex-ploration, recognize that there are boundto be shades of agreement as to what it all

means, and accept that an approach to aconsensus involves an evolution guided byutility rather than absolutes. Some of myown favorite quotations speak to the sub-ject: Einstein’s restatement of Occam’s Ra-zor: “Things should be kept as simple aspossible, but not simpler”; Chesterton’sdefinition of paradox as “Truth standingon her head to get attention”; and an ad-age from the early days of the chemios-motic battleground: “The heat of the con-troversy is inversely proportional to thehardness of the data.” This latter reflects ageneral observation that, if a hypothesissurvives the falsification test often enough,people stop arguing about it. Finally, Ioffer with some hesitation a remark of Pe-ter Mitchell’s: “The trouble with most sci-entists is not that they don’t have goodmemories, but that they don’t have goodforgetories.” I gathered from conversationthat what he meant was somethinglike the following: To be effective, a work-ing scientist must be selective in terms ofwhat data and interpretations he or shetakes seriously. Data is often conflicting,

and interpretations are often misleading.

The selection filters (sorting algorithms)

used are not just straightforward applica-

tions of logic. Instead they represent a so-

phisticated balancing act that must also

include subjective elements: personal

judgments of a colleague’s intellectual

worth, scientific integrity, disinterested-

ness, etc. We might set against this Richard

Feynman’s remark: “To guess what to keep

and what to throw away takes consider-

able skill. Actually it is probably merely a

matter of luck, but it looks as if it takes

considerable skill.” The messiness of sci-

ence is a reflection of the humanity of sci-

entists. The success of the enterprise comes

from the application of “reconciliation de-

vices” such as the perspective achieved

over time, summarized as the “Truth Will

Out Device (TWOD)” by Gilbert and

Mulkay [54]. Time allows the mechanism

of falsification to play itself out against new

experience: Popper’s idea in action. The dud

hypotheses are weeded out.


martial sophistication are not badtests of congruity with reality [35]. In awider view, the testing of life againstthe external thermodynamic world bysurvival of extant forms represents �4billion years of exploration of the na-ture of the real world in terms of phys-ical and chemical interactions thathave allowed the survivors to exploitthermodynamic potential. This explo-ration, and its more recent extensionthrough deliberate, conscious experi-ment, hypothesis testing, and exploita-tion of the artifacts of civilization, pro-vides an epistemological foundation ofsome security that should not be lightlydiscarded. Although the details are provi-sional, we can have some confidence thatthere’s a real world out there and that ithas followed invariant laws for those 4billion years (and longer, because ourknowledge of the early cosmos is basedon the same laws), even though our un-derstanding of those laws must be provi-sional.

UNIVERSAL DARWINISM—MEMES ASUNITS OF SEMANTIC CONTENTThe insubstantial nature of semanticcontent has obvious implications forany discussion of the mind; as notedearlier, consideration of similar in-substantialities in different forms andunder different names has driven phil-osophical speculation down the centu-ries. More recently, Dawkins [13, 36, 40]has introduced the notion that the basicbuilding blocks of knowledge are self-ish, mental agents that are transferredfrom mind to mind, and replicatethrough imitation, for which he coinedthe name memes [14]. The idea wasintroduced as the last chapter of TheSelfish Gene [13] and borrowed by asso-ciation the central tenets of that treat-ment. Dawkins’ brilliance makes theidea almost irresistible; it has an ap-pealing simplicity and recognizes anevolutionary potential that allows a use-ful extension of ideas from genetics andtheir application to cultural transmis-sion. The meme theme has been ex-ploited in a substantial literature (seefor instance Blackmore’s enjoyable ac-count [55] and reference therein). Den-nett [56], in a wide-ranging and insight-

ful exploration of consciousness, hasinvoked the idea of memes among a setof hypotheses to account for the pecu-liar properties of the human mind.

After a strong beginning, the fieldof memetics has fizzled; it is held backby lack of a cohesive idea of just whata meme is and by a disappointing lackof practical application. However,memes seem to show the same lack ofthermodynamic consequence as se-mantic content, and it seems produc-tive to view them simply as units ofsemantic content involved in informa-tion transmission at the cultural level.Neither Dawkins nor Dennett explic-itly recognized the insubstantial na-ture of memes, but the question oftheir nature cannot be adequately ad-dressed without such recognition. Ifmemes are units of semantic content,then it should be clear from the abovediscussion that they could only “in-fect” minds already prepared to de-code them. Simple “imitation,” nomatter how we stretch it, does not de-scribe this process adequately. With-out a mind and the history of its evo-lution, there would be no memes.Memes did not spring from nowhere:they originated in minds. They can betransmitted only through the agencyof thermodynamic carriers, the physi-cochemical mechanisms already dis-cussed above. They have value andsignificance only after translation andinterpretation and in the context of anappropriate education. We musttherefore view memetics against anevolutionary and an educational/so-cial background. This was at leastpartly recognized by Dawkins andDennett in their discussions of theevolutionary significance, the devel-opmental modeling through educa-tion, and the role of “mutation” ofmemes in cultural progress; a similartheme is developed further by Black-more [55]. However, all these discus-sions have been in the framework ofthe replicator role, the “selfish repli-cator” scenario, and the analogy withgenes. Without a commonfeature, thisanalogy is tenuous. A common themeis provided by the more general viewproposed here, which recognizes se-

mantic transmission as the element

binding these two replicators—indeed

as a fundamental aspect of Universal

Darwinism (Figure 13)—and suggests

that the nature of the evolutionary

process is determined by the need for

both the “engineering side,” which has

a well-defined physical context that I

have lumped under thermodynamic,

and semantic content, which intro-

duces no thermodynamic burden, has

value only in context, and requires

translational processing.

There are some simple memes that

behave as advertised: fads and fash-

ions or that commercial jingle you

can’t get out of your head; for exam-

ple—Dave Barry has called such men-

tal dross Brain Sludge. However, as

discussed in detail by Dennett [56],

the mental processes underlying con-

sciousness come in many variants,

from the innate functions on up.

These operate on a massively parallel

scale, with different time constants for

different sensory components, and

the level of complexity is further com-

pounded by interferences from psyche

and the hormones [15]. Both Dawkins

and Dennett certainly recognize dif-

ferent levels of complexity; both build

wider potential into the concept

through the mutability of memes and

by suggesting additional interactions

between memes that provide links to

the richer context of evolutionary

epistemology [37–39, 41], and Black-

more [55] explicitly recognizes the

debt to Popper. Nevertheless, some-

thing is missing. Perhaps the argu-

ment is distorted by the unfortunate

choice of “selfish” as a descriptor,

with its anthropocentric overtones,

and implications of individual choice.

Dawkins has expended some effort in

a revisionist shift to a position in

which the “selfish replicator” has be-

come a “selfish cooperator,” introduc-

ing a schizophrenic element to the

discussion, leading to confusion be-

tween several different properties,

which might be clarified by recogniz-

ing the role of semantic transmission

and its two components.


CONSCIOUSNESS: THE NATURE OFSEMANTIC CONTENT DOES NOTFORCE US TO A DUALISTIC VIEWDennett’s treatment of consciousnesscenters on the use of experimental datafrom cognitive neurosciences [57] to ex-plore the nature of consciousness. Aninitial objective is deflation of the ideaof an insubstantial Central Controlleroperating the curtains in a CartesianTheatre, a concept associated with thedualistic view of mind as immaterialand distinct from the substantial body.To achieve this, Dennett must use astrictly materialistic approach. Thetreatment depends on analysis of thecomplexity of processing, its highly par-allel nature, the different timescales ofconscious and subconscious process-ing, and the potential for “advanced”behavioral patterns inherent in thecombinatorial possibilities. The memesbecome part of a metaphorical bestiary

that serves as a replacement for theCentral Controller: memes of differentsorts, agents, homunculi and demons,all competing in a sort of Brueghelianhinterland underlying consciousness(Figure 14). Both here, and in his treat-ment of epiphenomena, Dennet fails tonotice the substitution of one immate-rial entity by another. If memes are todo their thing while retaining their in-substantial status, what about theseother ghosts? Do we have to resurrectdualism? Dennett does a nice job ofmaking a highly technical area accessi-ble to the outsider, and the resort tometaphor is well justified in this con-text. But “Consciousness Explained”? Itseems to me necessary to face up to theinsubstantial nature of semantic con-tent before considering the matter set-tled. I want therefore to re-emphasizehow the properties of semantic trans-mission impose constraints on mental

mechanisms. We can recognize the in-substantial nature of semantic contentwithout recourse to a traditional dual-istic view as long as we keep in mindthat, in addition to semantic content,semantic transmission requires a car-rier, encoding, and a translational ma-chinery, and that these “engineering”aspects have thermodynamic compo-nents, even if the meaning does not.These physical components tie the pro-cess to the real world. There is a dualityin play here, but not one that requiresus to abandon scientific enquiry. Thefeatures that led to the idea of a distinc-tion between mind and body are due tothe conjoined participation of two com-ponents in a single process.

The idea developed earlier, that allevolving systems must operatethrough semantic transmission, hasan obverse: only systems that areevolving have need of a semantic in-heritance. Only life has meaning;meaning only has significance in thecontext of evolving systems. However,the meaning—the semantic content—can only have expression in the con-text of a thermodynamic framework,can only be transmitted through themedium of a thermodynamic carrier,and can only be appreciated via atranslational machinery. Although theextraordinary features of the mind liein its ability to make manifest the in-substantial, there’s nothing unnaturalabout this. The semantic content ofgenes and memes share the same in-substantial property, so we could per-haps also marvel at how the pheno-type reflects the insubstantial messageof the genotype. We should recognizethat both translational feats are re-markable. Perhaps it is because weknow so much about the engineeringside and are so used to being our-selves the embodiment of the geneticmessage that we do not see this latterface of evolution as exceptional. Orperhaps it is because of the differencein linkage between the transfer of se-mantic content and the translationalmachinery needed for further matura-tion. The linkage is tight in reproduc-tion by cell division (the genotype iscopied, and half the somatic machin-

FIGURE 13

Scheme to show the role of semantic transmission in Darwinian evolution. All evolving systems follow thesame pattern in which the meaning of a thermodynamically encoded message is transmitted via atranslational machinery to allow reproduction. The overall process is subject to error, so that the translationalproduct may include “mutations.” The translational product is tested in an “ecosystem” against competitors.Products that survive reproduce, thus propagating the more successful traits, leading to improvement. Inreproduction (enclosed by dashed line), the step for semantic transmission is shown as separate from thatfor the copying of the replicator to accommodate obvious difference in the degree of linkage for differentforms. The tightness of linkage decreases in the order: bacteria reproduction by cell division; asexualreproduction in the protista; sexual reproduction in the metazoan; and supra-phenotypical inheritance. In allcases, the transmitted message needs to find itself in a competent translational environment. However, thelinkage between transfer of the message and the translational apparatus also depends on the caseconsidered, in a similar order (see text for explanation). The mature phenotype is reproduced after whatevermutation changes it, so that what looks like a cycle in this 2D representation would project to a 3D version(see Figure 12) and a spiral through time in 4D (see Figure 6). All processes that we perceive as cyclicalare spirals through time, with pitches and lengths spanning many orders of magnitude. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]


ery follows the copy) and becomeslooser in sexual reproduction (the fe-male gamete carries with it the so-matic machinery, but the male is justa chemical carrier of semantic con-tent) and separate in the supra-phenotypical case (memes require athermodynamic carrier, but depend onfinding a brain). To reinforce Dawkins’“viruses of the mind” idea [40], it’sworth noting that reproduction of vi-ruses follows a similar pattern: the se-mantic content is carried by the RNA orDNA, but the carrier has to find a hostwith a translational machinery.

KEEPING TRACK OF THEINSUBSTANTIALI have suggested above that the mindhas the ability to make manifest the in-substantial; it also has the ability to in-vest with meaning the spatiotemporalevolution of the environment. Theseseemingly miraculous functions are in asense quite primitive. We can dissect ata biochemical level the mechanismsthat provide unsophisticated versionsof these functions, for example in thebacterial response to gradients dis-cussed above. Competing biochemicalprocesses, with time constants for acti-vation, refreshment and decay, makepossible the detection of environmentalvariables represented by spatiotempo-ral gradients that determine behavior. Ittakes no great stretch of imagination tosee that similar principles might under-lie the more sophisticated behaviorsthat we recognize as consciousness.

Consciousness is usually discussed inan anthropocentric context. However, itseems likely that the underlying mecha-nisms of our perception machine arequite primitive and common at least to allvertebrates. Any organism that has theability to focus attention in a time-adap-tive fashion and relate behavior to asso-ciative memory has the need for the samemechanisms that drive our more sophis-ticated apparatus. The differences lie inthe sophistication of linkages to memory.In any case, the lack of thermodynamicconsequence of semantic content (or ofmemes) gives rise to a necessity for somerepresentation in “thermodynamic” formof any semantic or sensory component

that makes it through to the perceptionmachine. Since the semantic content isinaccessible to measurement, the moreimmediate task in trying to understandconsciousness is to take advantage of thetools available that explore the linkage tophysical or chemical processes. We cankeep track of this “invisible man” throughhis footprints in the snow, through obser-vation of the manifestations of this ther-modynamic linkage. Techniques likefMRI (functional magnetic resonance im-aging), EROS (event-related optical sens-ing; see Figure 15), and the electromag-netic recording of neuronal activity [58]take advantage of these physical manifes-tations.

While the initial translation of physi-cochemical input must be hard-wired,the upper echelons of our processing de-pend on learned associations retained inmemory; we recognize a dog because welearned about dogs from the family, orfrom Dick and Jane. The mechanisms ofmemory, either short or long term, arestill unknown, but whatever hardware isused, there has to be some physicochem-

ical linkage that can operate rapidly and

reversibly, both with the later stages of

translational processing of sensory input

and with the mechanisms that allows us

to be conscious of an external world.

These latter are also unknown, but are

likely related to the �40-Hz gamma

rhythms involved in higher mental activ-

ity. Since the focus of attention can shift

on this time scale, the slate must be con-

stantly refreshed and updated both from

memory and from sensory input. The

thermodynamically stable encoding of

meaning in the “perception machine” is

necessarily transitory. Because perception

is subject to modulation by memory, the

two must share common linkages; mem-

ory and sensory input must both have

connections that allow refreshment,

comparison, and updating of the appara-

tus for consciousness. From the timescale, the entire apparatus must functionin the time domain of cell excitability, sothat even if long-term memory involvesprotein synthesis or cellular remodeling,it must also be accessible to those rapid,reversible linkages. Because the basic ma-chinery of perception is relatively primi-tive, it seems likely that it is seated in arelatively primitive region of the brain,perhaps the thalamus rather than thecortex [58–60]. Since the whole set ofmechanisms is realized in a highly paral-

FIGURE 14

Detail from “The Fall of the Rebel Angels” by Pieter Brueghel the Elder, from the Webmuseum, Paris(http://www.ibiblio.org/wm/paint/auth/bruegel/). Dennett’s treatment of the states underlying con-sciousness [56] involves demons, homunculi, memes, and other agents, a metaphor nicely illustratedby Brueghel’s painting. [Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]

Consciousness is usuallydiscussed in an anthropocentric

context.


lel, spatially delocalized architecture, thelinkages to memory and the cortex mustbe governed by local preprocessing ele-ments associated with local memorymodules. There is no reason why theseshould not be quite widely dispersed onan anatomical scale, but, in order to bepart of a coherent mechanism of percep-tion, they do need to share commoncommunication channels and to be con-trolled by a common synchronization.

CONSCIOUSNESS OF SELFThe reversible-linkage mechanism out-lined above offers a “black-box” solutionto the “insubstantial” nature of con-sciousness, a highly parallel set of pro-cessing functions that link memory andsensory input to perception of an internal“image” in a mechanism that is “re-freshed” by editing and updating func-tions, perhaps through mediation of thegamma rhythms. But this still fails to ex-plain our perception of meaning or theillusion of self.

Dawkins’ introduction of the idea ofmemes in the last chapter of The SelfishGene [13] was so tongue-in-cheek, thatSusan Blackmore [55] at her first reading

“dismissed the idea . . . as no more than abit of fun.” The simple message on a pre-liminary reading seemed to be that indi-vidual memes have a purpose of theirown that is independent of (and can sub-vert) that of the mind. This view does notdo justice either to the complexities ofthought or to the extent to which we de-termine our thought processes. Indeed,toward the end of the chapter, Dawkinsrecognized this in a discussion of how“our conscious foresight—our capacity tosimulate the future in imagination—could save us from the worst selfish ex-cesses of the blind replicators” [13]. Inlater publications [22, 61], Dawkins hasemphasized the selfish cooperator face ofthe gene analog and developed the moresubtle point that, after any changethrough mutation, the gene has to survivein the context a variable set of gene bud-dies, which are exchanged among a pop-ulation. Selection pressure occurs at sev-eral levels. An essential requirement isefficient integration of the gene productinto a metabolic machine, because, with-out an efficient functional whole, any “lo-cal” advantage conferred by a particularmutation would be worthless. Selection

in the context of an external ecologyworks on this functional whole, but nowin the context of a population. Transfer-ring this idea to the meme makes for anevolution of memes in the context both ofcohabitation in a meme population in theindividual mind and of an external cul-ture, so that the fate of an individualmeme is subsumed under the survival ofa composite whole. We might think of themind as a dynamic aggregate of thesecomposite memes, integrated into agreater whole that we experienced as self.Blackmore [55] suggests that this sense ofself is an illusion and that we are at themercy of the memes; “we” are ourmemes, no more, no less—an echo ofHume’s view of the self as just a bundle ofperceptions. Dennett recognizes that theJoycean virtual machine of his model ofconsciousness would be a zombie unlessit had evolved with the potential for adap-tive behavior, mutation of ideas, a mech-anism for distilling meaning from theBrueghelian subconscious, and a per-sonal and social identity.

As I see it, at some level we need torecognize that there is a willfulness tomental processing, a sense of self rein-forced by its survival potential. Popper[38] had earlier suggested that our abilityto manipulate models in the mind allowsus to discard unworkable ones withoutgetting killed for them—“Good tests killflawed theories; we remain alive to guessagain”—an idea echoed in Dawkins’ re-mark above. We seek out informationand informed sources. Saul Bellow’scharacter in The Dean’s December [62]has a nice remark to the effect that helikes living in the mid-West because“. . . By the time the latest ideas reachChicago they’re worn thin and easy to seethrough.” The sophistication of our sen-sory processing depends absolutely onwhat we have learned from life; Bellow’snovel is, at one level, an exploration ofthis theme (as indeed is much of ourgreatest literature). Organized thought isan extended set of coupled processes, inwhich different components, facts, ideas,filters, Kantian maxims, Campbell’s vic-ars, etc., play different mechanistic roles.New facts and old ideas are reshuffledand new relationships are tested in themind using sorting algorithms; we talk of

FIGURE 15

EROS maps showing cortical neuronal activity measured noninvasively through changes in nearinfrared transmission. The optical apparatus measures IR transmission through the brain cortex fromoutside the skull. Here, the stimulus was a thumb-twitch, elicited by electrical stimulation of the wrist,and maps are similar for different repetition frequency (top). The responses occur at 16–32 ms afterstimulation. Events on this time scale are obviously registered (since they induce the response seen),but are probably not yet consciously perceived. The Z score represents differences from baseline. Thedarker gray shows the area monitored. The brain image was from structural MRI, used to determinethe brain area corresponding to the response. Similar maps can be measured following sensorystimulation by sight or sound, but these elicit responses with greater latency, and in areas dependent onsensory and semantic context. The time scale of measurement here is fast enough for exploration of themechanisms underlying consciousness (from Maclin et al. [102]; reprinted with permission from Elsevier).


“weighing the facts” and an “informedmind.” These adaptive filters are not ge-netically linked. They are learned, neces-sarily in an incremental fashion. We donot expect a child to have the same dis-crimination as an adult. The sorting aidsare just as much a part of the semanticinheritance as the facts and ideas in-volved in the shuffling process; they arethe very stuff of our education in its widermeaning. As Dennett [56] notes, a suffi-ciently complex virtual machine could beimagined to explain much of our mentalcomplexity. However, the lumping of allaspects of this complexity under thememetic heading seems to me to be over-stretching the idea. We could of courserecognize different classes of meme: ideamemes, filter memes, memory memes,vicar memes, etc., but then we wouldquickly reach a condition in which theterm meme and the concept become re-dundant. Indeed, Dennett’s resort tometaphor seems a recognition of this di-lemma. Without wishing to detract fromthe utility of memetics and the fruitful-ness of the idea, the features that makesense can be subsumed under the insub-stantial nature of all semantic content.The accusation, also leveled against Pop-per’s three-world model, that this resur-rects duality, can be defused, as notedalready, by a recognition that the seman-tic component is necessarily coupled tothe “engineering” side.19

The sorting tools available includethe rules of grammar and syntax, math-

ematics, and logic; the physical laws;moral laws like the Golden Rule; thedemocratic ideals that in part derivefrom it; and belief systems. The internalmental process is iterative—a sortingand rearranging that is to some extentunconscious—with intermediate hy-potheses tested against an evolvingmodel maintained in the mind, appro-priate to the area under consideration.The important point is that as our edu-cation evolves, we are more and morecapable of making a choice, includingthat of which sorting aids to use, and weexercise our critical faculties. Howeveramorphous my “self” is in terms of an-atomical locality, “I” appear to directmy own focus of attention. Clearly, theself has an unhappy tendency to mud-dle things up; Blackmore [55] has a nicechapter on subjugating memes by med-itation, very much along the lines ofBuddhist practice, and this might wellprovide a mechanism for release fromself, with therapeutic benefits (to re-duce nirvana to the prosaic). However,abdication of self is not a proven recipefor a successful society. Rather than as-cribing the illusion of self to the com-plex interactions between squabblingmemes in a meme playground, it seemsto me more productive to view the mindas a meme stud farm. Then I can go outand rustle up that pretty memeticsmare and sire her with that strappingShannon stallion, and pretty soon I’llhave a fine bunch of frisky semanticfoals that others might be interested in.Rather than a wimpy acquiescence inthe dominance of memes, let’s corralthem and lick them into a self that wecan be proud of! That way, we can ex-pect to be held responsible for our deci-sions, expect others to be responsible fortheirs, and maybe get back a society thatvalues responsibility rather than excuse.

“SORTING AIDS” AND THE SURVIVALOF SOCIETIESMy third quibble with the simplememetic view is in the evolutionarymode. As originally expressed, thisborrowed from Dawkins’ selfish genehypothesis the idea that the replicator(gene or meme) was the driving agent.Of course, it is clear in each case that

the semantic content is the compo-

nent of information transmission

passed between successive genera-

tions and that its thermodynamic rep-

resentation is therefore necessarily

the entity that endures, the point that

I think Dawkins intended. Taking the

view that information transfer re-

quires both message and vehicle and

recognizing that the former is insub-

stantial propels us to a perspective in

which the replicator is inseparable

from its thermodynamic context, and

its evolution is inseparable from an

ecological context. For the meme, this

surely involves a testing of man and

his mind and the complex ideas it pro-

duces, against the world experienced

as real, rather than survival of any self-

ish meme. The emphasis has to be

shifted to the more sophisticated ver-

sion of the meme as a composite

whole. Along these lines, Dawkins and

Dennett both integrate the fate of

memes with that of their hosts (a

meme that leads its bearer to disaster

won’t last long in the meme pool), but

this leaves one wondering if there is

any sense in imagining the meme as

an “independent” evolutionary entity,

as seems to be championed by Black-

more. Certainly (unless we want to ex-

plore supernatural possibilities), no

meme can exist independently of its

thermodynamic context, and an ex-

tended thermodynamic matrix of in-

teracting entities (an external world,

people and their minds, books, com-

puters, etc.) is needed for meme ex-

change, selection, and conservation.

The ecological niches for meme com-

petition are the individual mind and

societies. But this idea is no more than

Popper’s three worlds reformulated.

We offer our thoughts to “the

world” for a wider testing against “re-

ality.” That the reality includes a so-

cial and educational context that

might be hostile, wrong, or muddled

does not detract from the willful na-

ture of the process. However, the na-

ture of our local reality will certainly

affect the outcome. The particular cul-

ture has consequences for the individ-

ual; poor Bruno at the stake comes to

19This idea is not new. As Parmenidesputs it, “Therefore thinking, and that byreason of which thought exists, are oneand the same thing, for thou wilt notfind thinking without the being fromwhich it receives its name” [96]. Despitethe general perception of his role as themodern source of the dualistic view, thesame idea can even be recognized inDescartes’ writings. He discussed theclose and intimate union between mindand body. His observation that matteroccupies space but that thought does notmight be seen as an early version of mysuggestion that semantic content cannotbe evaluated through physical measure-ment.


mind.20 In the longer terms, choice ofsorting aids also has consequences for thesurvival of the culture sponsoring them. Acivilization or social group is defined asmuch by the sorting aids it condones asby anything. The sorting rules of any par-ticular society have the status of, for ex-ample, the rule books for Wittgenstein’s“language games,” or similar scenariosfrom justificational epistemologies; theydetermine what is acceptable to the club.Choice of sorting aids on both the indi-vidual and societal sides is at the heart ofthe social contract.

It seems to me no accident that themost successful exploitation of physicaland intellectual resources in the modernworld has occurred in the context of dem-ocratic societies. The continuous applica-tion of human endeavor needed to main-tain our highly unstable complexsocieties is best served through the will-ing participation of its members. “Will-ingness” can be engendered through dif-ferent mechanisms; religion, nationalism,and totalitarian fervor have obviouslyworked in the past. But—and this is theview championed by Popper—in demo-cratic societies, the feedback loops be-tween government and the people, andfreedom to choose, open up possibilitiesfor error correction that seem to have alonger term advantage. An importantcomponent of the struggle for democracyhas been the idea that people put moreeffort into tasks from which they receive abenefit. This implies choice, and hencethe explicit recognition of liberties, re-sponsibilities, and, as famously ex-pounded in the Declaration of Indepen-dence, the “pursuit of happiness.”Democratic mechanisms allow for evolu-tion of the sorting rules condoned by so-ciety. In contrast, societies based on fun-

damentalist religious or totalitarianprinciples, in which new ideas have topass a litmus test against dogma, aredoomed to the same stagnation as per-vaded the Middle Ages. The sorting toolsinculcated at the madrassah, or at BobJones University,21 are not designed tobroaden either the limits of experience orthe chronognostic range (Figure 16).

STANDING ON THE SHOULDERS OFGIANTSIt is in the context of perception that Iwant to return to the chronognostictheme introduced above. The graph ofFigure 17 can be thought of as a plot ofsubjective time against objective time.Our objective measurement of time iscalibrated to the ticking of the cosmicclockwork. The “arrow of time” onthat scale is changed not a jot by ourinsignificant meddling. Our subjectiveexperience is more elastic. The initialslope of the graph represents themodest stretching of perception as thebiosphere evolved, linked to the in-creased sophistication in the entropicdelay introduced as complexity in-creased. The change in slope repre-sents a more dramatic stretching ofsubjective time resulting from ourevolution in the supra-phenotypicaldomain. A similar graph could havebeen presented showing the increasein effective force available for manip-ulation of the environment and wouldshow a similar change in slope. Thecorrelation between these two trendshighlights the link between chronog-nostic range and the power we wield,but the slope begs the question ofwhere we’re headed.

We take our perception of time somuch for granted that its implications be-

come invisible, so much so that even inthe discussions of evolutionary episte-mology, its role is largely implicit (thoughPopper gives it some prominence [37,38]). However, it is only in the context ofchronognosis that behavior has meaning;only through congruity with a polarity oftime that spatiotemporal behavior makessense. Life and evolution represent an ex-ploration of the nature of the physicalworld in which predictive success has de-termined survival.

Recalling Olsen’s coin-tossing “mira-cle,” it is pertinent to note that two viewsof the same experience lead to differentevaluations. From a statistical perspec-tive, the outcome was prosaic. From thetemporal perspective, the sequence wasextraordinary. We can view life fromthese two perspectives: from the statisti-cal, we get the Ronald Reagan view ofbiology: “If you’ve seen one redwood,you’ve seen them all” and from the evo-lutionary, every living creature is theunique fruition of a 4-billion-year exper-iment. Chronognosis is the most imme-diate sorting agent through which a newform or new hypothesis is tested againstthe physical world. This view of temporalperception seems almost tautological, be-ing built into the whole concept of causeand effect. Indeed, for Kant, the percep-tion of time was a priori, obtained directlyfrom intuition, a view that seems wellsupported by the recent work on biolog-ical clocks. At the molecular level, evolu-tionary mechanism is predicated on tem-poral polarity: the central role of sequencein allowing an ordering of base pairs, or ofamino acids, which underlies combinato-rial amplification and information pro-cessing, and selection rules that allow evo-lution to occur; the very ideas and thephysical realities they represent are im-

20Giordano Bruno was burned at the stakefor his support of the Copernican hypoth-esis, including his extension of the uni-verse beyond the “immobile sphere” of thefirmament (to give us a view of the stars asother suns with their own planets, people,and gods) and for his suggestion that theCreator could not be less than this largercreation. “On January 20 of the first year ofthe new century (1600) Pope Clement VIII

ordered that Bruno be handed over to sec-ular authorities for punishment. Sentenceis read on February 8th: Bruno is declaredan impenitent, obstinate and pertinaciousheretic, to be stripped of priesthood andexpelled from the Church, and his worksto be publicly burned at the steps of St.Peters and placed on the Index of Forbid-den Books. Bruno replies: �Perhaps youwho pronounce my sentence are in greater

fear than I who receive it.’ ” The secularauthorities burned Bruno as well (takenfrom http://members.aol.com/Pantheism0/brunlife.htm).21This from the BJU Web page: “The Op-portunity Place: A liberal arts, non-de-nominational Christian university, BJUstands without apology for the old-timereligion and the absolute authority of theBible.”


bedded in time. As noted already above,the temporal context of life and evolutionmeans that the combinatorial possibili-ties at the metabolic and ecological levelsare greatly expanded by the increased di-mensionality. At the primitive level, abacterium won’t reach its food if it mis-reads the gradient. Similarly, at higherlevels, the everyday experience of time iswhat compels all action and reaction. Atthe human level, our sophistication en-ables us to entertain philosophical para-doxes around time (Figure 17), but doesnot relieve us of the “tyranny of time” incontrolling our daily lives. Our whole ap-paratus of perception is based on ourchronognosis, and this linkage operatesall the way from the innate to the fullflowering of intellect. However, it is ourability to anticipate the future more suc-cessfully than other species that has ledboth to our present success and to ourpresent peril as unthinking wielders ofpower.

The development of a modern view oftime and space, including other aspectsof evolution and the “biological arrow oftime,” has been discussed with insight bymany others [17, 30, 31, 33, 35–39, 41,63–69].I will not embellish my own dis-cussion in these areas further, except tonote that our appreciation of time en-gages the highest levels of our under-standing. The brief history of time thatemerges from the biological perspectiveis in a sense more inclusive than thatfrom the cosmological view. Thanks tothese authors and their kind, what weknow of the cosmos and of time arenow parts of the common semantic her-itage. The mind of a Panini, a Socrates,a Dante, a Bacon, a Shakespeare, or an

22Albert Einstein’s (1934) address to agroup of children: “My dear children: Irejoice to see you before me today, happyyouth of a sunny and fortunate land. Bearin mind that the wonderful things thatyou learn in your schools are the work of

many generations, produced by enthusias-tic effort and infinite labor in every coun-try of the world. All this is put into yourhands as your inheritance in order thatyou may receive it, honor it, and add to it,and one day faithfully hand it on to your

children. Thus do we mortals achieve im-mortality in the permanent things whichwe create in common. If you always keepthat in mind you will find meaning in lifeand work and acquire the right attitudetowards other nations and ages.”

FIGURE 16

The relation between chronognostic range and evolution of the biosphere. The plot of chronognosticrange (CR, in years) against years before the present (YBP) is shown on a log vs. log scale toaccommodate the many orders of magnitude involved. The slope in the early part of the graph reflectsthe contribution from advances in behavioral complexity. A few “markers” will illustrate the approach.The bacterial world was constrained by the biochemical processes discussed in the text (CR in theseconds range; see also Figure 3). As evolution led to more advanced forms, the temporal perceptionlengthened, perhaps to the lifetime for an advanced vertebrate. But even in the case of early humansocieties, the range was constrained to a few 100 years and was limited by oral transmission. Withthe development of archival information storage, this was extended into the thousands of years range.The change in slope reflects an acceleration arising from the development of a supra-phenotypicalsemantic heritage. The choice of points is biased, since the latter part of the plot is based ondevelopments in the civilization of Europe and the West. However, in Western thought, it was not untilthe Copernican revolution, Bruno, and Galileo that our perception of the universe was stretched furtherthan the Greco-Roman-Arabic Ptolemaic tradition, and not until Darwin that our temporal range wasstretched beyond the “biblical limit.” In modern times, this range is determined by cosmological timescales, accessible to interpretation though advances in our understanding of astronomy and itsrelativistic scales of time and distance. The graph reveals a dramatic change in slope in the mid-13thcentury. A strong case is made by Jack Weatherford in Genghis Khan and the Making of the ModernWorld [103] that the flowering of intellectual life in the 14th century owed much to the spread of ideasafter the establishment of the Mongol Empire. This opened trade routes that facilitated an exchangeof ideas and culture between Europe, the Islamic Empires, and East Asia. Subsequent developmentsleading to the astonishing slope must include the dawn of humanism, the reformation, the develop-ment of literature in the vernacular, the renaissance, the age of exploration, the spread of printing inthe West, and the emergence of a rational science from the constraints of dogma. [Color figure canbe viewed in the online issue, which is available at www.interscience.wiley.com.]

We take our perception of timeso much for granted that its

implications become invisible . . .


Einstein may be extinguished, but bitsof each have spilled out into the exter-nal semantic inheritance to help us onthe road to a deeper understanding.They are immortalized in the widercontext of our society—their souls gomarching on.22 These are the giantswhose shoulders we stand on.23 Our ex-tensions of the limits of experience andof our chronognostic range are all partand parcel of the same thing: the evo-lution of the biosphere to take advan-tage of the thermodynamic opportuni-ties available, through refinement of thesemantic heritage, and increased com-binatorial complexity. Within thisframework, we have developed an un-derstanding of our universe as havingan independent physical existence inspace, governed by time-independent

laws.24 We have developed formal de-scriptions through which we can inter-pret the direction of time and the tem-poral behavior of matter in a way thathas allowed us to exploit our environ-ment to further our own advantage. Aswith all such models, these views areconjectural and will change, but itseems likely that the mechanism of atesting against the realities of the ther-modynamic world will continue to pro-vide a pathway for further refinement.

POSTSCRIPTThe reader who has borne with me so farwill not need to be reminded that withextension of the limits of our experiencecomes a linked responsibility. We can nolonger invoke religion in assigning re-sponsibility for our behavior with respect

to those aspects of our interactions withthe natural world for which we claim anunderstanding. A remarkable feature thatemerges from the above discussion is theimpact that the essentially biologicalfunctions represented by the extensionsof human behavior in the supra-pheno-typic domain have had on the very sys-tem that has generated them. The ex-traordinary developments of the pastcentury accompanying the advances inscience, computation and technology,and their exploitation in the service ofhuman kind have left us in an unprece-dented situation. The “effortless” seman-tic content, shared as the collective wis-dom of our civilizations, is in starkcontrast to the way in which it has en-abled our species to direct energy into aremodeling of this world. This in turn has

23Bernard of Chartres (d. c.1130): “We arelike dwarfs on the shoulders of giants, sothat we can see more than they, and thingsat a greater distance, not by virtue of anysharpness of sight on our part, or anyphysical distinction, but because we arecarried high and raised up by their giantsize.” Quoted by John of Salisbury in TheMetalogicon (1159; from [97]). This fa-mous metaphor was also used by IsaacNewton in a letter to Robert Hooke (1676).There is a sneaking suspicion, because of

the hostility between the two and Hooke’ssmall stature, that the sweeping vision ofBernard’s phrase was not all that Newtonhad in mind.24The view of time in modern physics is tiedto controversial issues in both cosmologyand quantum mechanics. In terms of thelocal universe, the constraints of conven-tional physics and relativity set well-definedlimits on measurements through which ahypothesis can be tested. However, some in-terpretations of quantum mechanics (and

its extension to string theory), and some cos-mological theories, include multiworld/multiuniverse models in which the presentimpossibility of measurement limits theepistemological certainties. Hypotheses arethen so much less constrained (as is wellrecognized by some in the physics commu-nity [98]) that one wonders if they wouldpass the tests implicit in the Jones ruling (seefootnote 13). See http://www.life.uiuc.edu/crofts/papers/Epistemology_and_QM.pdf fora naive biologist’s perspective.

FIGURE 17

Paradoxes of time: meme selection at the machine level? A small sample of the images delivered by Google in response to arrow of time; the recall of associativeimages from the Web represents a second-hand dipping into a pool of images deposited by human agency. Left: “Time flies like an arrow”, and middle: “Fruitflies like a banana”, both from Harvey Galleries (www.harveygallery.com/, reproduced with kind permission of Henry Harvey); right: the CPLear experiment(www.cern.ch/cplear/) to test charge, parity, time (CPT) invariance, reproduced with kind permission from CERN. [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com]


put an obvious strain on our relation to therest of the biosphere and our physical en-vironment. Over the past few decades, ourmeddling has changed not just the bio-sphere, but the whole balance of energyfluxes through which the biosphere ismaintained. Ozone depletion from the cat-alytic effects of halogenated hydrocarbonswas a near-catastrophe, which, despite theinternational response, will be with us fordecades. Global warming from the in-creased levels of CO2 accompanying burn-ing of hydrocarbon reserves will have pro-found effects on ecology and is alreadymelting ice-caps and raising the levels andincreasing the acidity of our ocean. In theUnited States, these warning signs havebeen passed off by ignorant politicians asinconsequential,25 so that the profligateconsumption continues unabated. Perhaps

if we pause to reflect on the intimacy of ourrelationship with the biosphere, the deli-cate balance with the atmosphere andoceans, and the fragility of our highly or-dered world, we might recall ourselves torecognition of our responsibilities. No onewill preserve our world if we don’t.

AcknowledgmentsI am grateful to Gordon Baym for helpingme to recognition of the difficulties inher-ent in informational entropy, to TonyLegget for pointing me to Huw Price’snice book about the Arrow of Time, toGary Olsen for sharing with me his mira-cle, and to all three for interesting

and stimulating discussion in the earlydays of my interest. Special thanks also goto Bruce Hannon and to Bob Jones forextensive discussion and encouragement,to Michael Weissman, Tony Leggett andMartin Gruebele for their insights intoquantum mechanics, to Nigel Goldenfeldfor educating me on the evolution of thegenetic code, and to Gabriele Gratton andMonica Fabiana for many interesting dis-cussions on neurosciences and cognition.All corrected faulty aspects of my think-ing, but I take full responsibility for anyremaining errors. I’d also like to thank mywife Chris Yerkes for her valuable input,restraints on my wilder flights of fancy,and for her support, and to my childrenand their friends for their interest. Sup-port for work on bioenergetic aspectsfrom the National Institutes of Health,National Institute of General MedicalScience grant GM35438 is gratefullyacknowledged. A preliminary version ofthis material was presented at a PacificInstitute for Theoretic Physics meetingon Emergence (http://pitp.physics.ubc.ca/archives/CWSS/emergence/) inMay 2005.

REFERENCES1. Broda, E.; Ludwig Boltzmann-Man-Physicist-Philosopher; Oxbow Press: Woodbridge, 1983.2. Crofts, A.R.; Wraight, C.A.; Fleischman, D.E. Energy conservation in the photochemical reactions of photosynthesis and its relation to delayed fluorescence.

FEBS Lett 1971, 15, 89–100.3. Lavergne, J. Commentary on: Photosynthesis and negative entropy production, by Jennings and coworkers. Biochim Biophys Acta 2006, 1757, 1453–1459.4. Nemani, R.R.; Keeling, C.D.; Hashimoto, H.; Jolly, W.M.; Piper, S.C.; Tucker, C.J.; Myrneni, R.B.; Running, S.W. Climate-driven increases in global terrestrial

net primary production from 1982 to 1999. Science 2003, 300, 1560–1563.5. Running, S.W.; Nemani, R.R.; Heinisch, F.A.; Zhao, M.; Reeves, M.; Hashimoto, H. A continuous satellite-derived measure of global terrestrial primary

productivity. Bioscience 2004, 54, 547–560.6. Mitchell, P. Coupling of phosphorylation to electron and hydrogen transfer by a chemi-osmotic type of mechanism. Nature 1961, 191, 144–148.7. Mitchell, P. Chemiosmotic coupling in oxidative and photosynthetic phosphorylation. Biol Rev Cambridge Phil Soc 1966, 41, 445–502.8. Saier, M.H. Peter Mitchell and the Vital Force: http://www-biology.ucsd.edu/�msaier/transport/petermitchell/MitchellFrame-1.html, 2005.9. Harold, F.M. Gleanings of a chemiosmotic eye. Bioessays 2001, 23, 848–855.

10. Prebble, J.; Weber, B. Wandering in the Gardens of the Mind: Peter Mitchell and the Making of Glynn; Oxford University Press: New York, 2003.11. Saraste, M. Oxidative Phosphorylation at the fin de siecle. Science 1999, 283, 1488–1493.12. Shannon, C.E. A mathematical theory of communication. Bell System Techn 1948, 27, 379–423.13. Dawkins, R. The Selfish Gene; Oxford University Press: New York, 1976.14. Dawkins, R. River out of Eden: A Darwinian View of Life; BasicBooks, The Perseus Books Group: New York, 1995; p 106.15. Hayes, B. The invention of the genetic code. Am Sci 1998, 86, 8–14.16. Vetsigian, K.; Woese, C.; Goldenfeld, N. Collective evolution and the genetic code. Proc Natl Acad Sci USA 2006, 103, 10696–10701.17. Dawkins, R. The Ancestor’s Tale: A Pilgrimage to the Dawn of Evolution; Houghton Mifflin Company: New York, 2004.18. Trewavas, A. Aspects of plant intelligence. Annals of Botany 2003, 92, 1–20.19. Allada, R.; Emery, P.; Takahashi, J.S.; Rosbash, M. Stopping time: The genetics of fly and mouse circadian clocks. Ann Rev Neurosci 2001, 24, 1091–1119.20. Alena, S.; Zdenka, B.; Martin, S.; Rehab, E.H.; Kristyna, L.; Zuzana, J.; Helena, I. Setting the biological time in central and peripheral clocks during ontogenesis.

FEBS Lett 2006, 580, 2836–2842.21. Bergson, H. Creative Evolution; Dover: New York, 1911.22. Dawkins, R. River out of Eden: A Darwinian View of Life; BasicBooks, The Perseus Books Group: New York, 1995.23. Eigen, M. Self-organization of matter and the evolution of biological macromolecules. Naturwissenschaften 1971, 58, 465–532.24. Eigen, M.; McCaskill, J.; Schuster, P. Molecular quasi-species. J Phys Chem 1988, 92, 6881–6891.25. Eigen, M.; Schuster, P. Stages of emerging life—five principles of early organization. J Mol Evol 1982, 19, 47–61.26. Davies, P.C.W. Emergent biological principles and the computational resources of the universe. Complexity 2004, 10, 11–16.

25One is tempted to observe that somepoliticians behave as if they have thechronognostic range of hollyhocks. Thespinless kowtowing of the Bush admin-istration and its supporters in Congressto the rapacious demands of commercial

interest, especially in energy policies thatexacerbate the effects of global warming,is likely to haunt us far into the future.Diamond [63] in Collapse underlineshow a limited chronognostic range canspawn bad policy. The estimate of chro-nognostic range (above) might be an in-sult to the biennial hollyhock; Diamondhas an account of the priorities of theearly Bush administration that suggestsa range of just 90 days.


27. Levinthal, C. How to fold graciously. In: Mossbauer Spectroscopy in Biological Systems; Debrunner, J.T.P., Munck, E., Eds.; Allerton House: Monticello, IL,University of Illinois Press, 1969; pp 22–24.

28. Woese, C. The universal ancestor. Proc Natl Acad Sci USA 1998, 95, 6854–6859.29. Woese, C. A new biology for a new century. Microbiol Mol Biol Rev 2004, 68, 173–186.30. Price, H. Time’s Arrow and Archimedes’ Point; Oxford University Press, 1996.31. Price, H. On the origins of the arrow of time: Why there is still a puzzle about the low entropy past. In: Contemporary Debates in the Philosophy of Science;

Hitchcock, C., Ed.; Blackwell, 2004.32. Katchalsky, A.; Curran, P.F. Non-equilibrium Thermodynamics in Biophysics; Harvard University Press: Cambridge, MA, 1965.33. Prigogine, I.; Stenger, I. Order out of Chaos: Man’s New Dialogue with Nature; Bantam Books: New York, 1984.34. Putnam, H. Words and Life; Harvard University Press: Cambridge, MA, 1994.35. Diamond, J. Guns, Germs, and Steel: The Fates of Human Societies; Norton, W.W. & Company, Inc.: New York, 1999.36. Dawkins, R. The Extended Phenotype; W.H. Freeman: Oxford, 1982.37. Popper, K.R. Objective Knowledge: An Evolutionary Approach; Clarendon Press: Oxford, 1972.38. Popper, K.R. Knowledge and the Mind-Body Problem; Routledge, 1993.39. Campbell, D.T. Evolutionary epistemology. In: The Philosophy of Karl Popper; Schilpp, P.A., Ed.; Open Court Publish.: La Salle, IL, 1974, pp 413–463.40. Dawkins, R. Viruses of the mind. In: Dennett and his Critics; Blackwell, 1993, pp 13–27.41. Bartley, I.W.W. Philosophy of biology versus philosophy of physics. In: Evolutionary Epistemology, Theory of Rationality, and the Sociology of Knowledge;

Radnitzky, G.; Bartley, I.W.W., Eds.; Open Court: La Salle, IL, 1987.42. Heylighen, F.; Campbell, D.T. Selection of organizations at the social level. In: The Quantum of Evolution: Toward a Theory of Metasystem Transitions;

Hehlighen, F., Joslyn, C., Turchin, I., Eds.; World Futures: The Journal of General Evolution, 1995.43. Kenny, A. The Wittgenstein Reader; Blackwell Publishing Ltd.: Oxford, 1994.44. Popper, K.R. Conjectures and Refutations: The Growth of Scientific Knowledge; Routledge: London, 1963.45. Latakos, I. The Methodology of Scientific Research Programmes; Worrall, J.a.C.G., Ed.; Cambridge University Press: Cambridge, 1970.46. Feyerabend, P. Against Method; New Left Books: London, 1975.47. Tarski, A. Logic, Semantics, Metamathematics, Papers from 1923–1938; Hackett Publishing Co.: Indianapolis, IN, 1983.48. Kuhn, T.S. The Structure of Scientific Revolutions, 2nd ed.; University of Chicago Press: Chicago, 1970.49. Miller, D.W. Critical Rationalism; Open Court: Chicago and La Salle, 1994.50. Bloor, D. Knowledge and Social Imagery, 2nd Ed. University of Chicago Press: Chicago, IL, 1991.51. Fine, A. Science made up: constructivist sociology of scientific knowledge. In: The Disunity of Science: Boundaries, Contexts, and Power; Galison, P., Stump,

D., Eds.; Stanford University Press, 1996, pp 231–254.52. Rorty, R. Philosophy and the Mirror of Nature; Princeton University Press: Princeton, NJ, 1979.53. Suskind, R. Without a Doubt. In: New York Times Magazine, October 17, 2004.54. Gilbert, G.N.; Mulkay, M.J. Opening Pandora’s Box: A Sociological Analysis of Scientists’ Discourse; Cambridge University Press: Cambridge, UK, 1984.55. Blackmore, S. The Meme Machine; Oxford University Press, 1999.56. Dennett, D.C. Consciousness Explained; Little, Brown & Co.: Boston, MA, 1991.57. Gazzaniga, M.S.; Ivry, R.B.; Mangun, G.R. Cognitive Neuroscience, 2nd Ed.; W.W. Norton & Co.: New York, 2002.58. Baars, B.; Newman, J.; Taylor, J. Neuronal mechanisms of consciousness: A relational global workspace framework. In: Toward a Science of Consciousness

II: The Second Tucson Discussions and Debates; Hameroff, S., Kaszniak, A., Laukes, J., Eds.; MIT Press: Boston, MA, 1998; pp 269–278.59. Crick, F.; Koch, C. A framework for consciousness. Nat Neurosci 2003, 6, 119–126.60. Green, A. The Science and Philosophy of Consciousness. http://www.users.globalnet.co.uk/�lka/conz.htm, 2003.61. Dawkins, R. Unweaving the Rainbow: Science, Delusion and the Appetite for Wonder; Houghton Mifflin: Boston, 1998.62. Bellow, S. The Dean’s December; Harper & Row, 1982.63. Diamond, J. Collapse: How Societies Choose to Fail or Succeed. Viking Penguin: London, 2005.64. Gould, S.J. Ontogeny and Phylogeny; Belknap Press: Harvard, MA, 1977.65. Gould, S.J. The Panda’s Thumb: More Reflections in Natural History; W.W. Norton & Company, Inc.: New York, 1992.66. Hawking, S. A Brief History of Time; Bantam Books: New York, 1996.67. Hawking, S. The Theory of Everything: The Origin and Fate of the Universe; New Millenium Press, 2002.68. Penrose, R. Singularities and time-asymmetry. In: General Relativity: An Einstein Centenary; Hawking, S.W., Israel, W., Eds.; Cambridge University Press:

Cambridge, UK, 1979, pp 581–638.69. Penrose, R. The Emperor’s New Mind: Concerning Computers, Minds, and the Laws of Physics. Penguin Books: New York, 1991.70. Petz, D. Entropy, von Neumann and the von Neumann entropy. In: John von Neumann and the Foundations of Quantum Physics; Redei, M.a.S.M., Ed.; Kluwer:

Amsterdam, 2001.71. Weaver, W. Introduction. In: The Mathematical Theory of Communication; Shannon, C.E., Weaver, W., Eds.; University of Illinois Press: Urbana-Champaign, 1948.72. Weinberger, E.D. A theory of pragmatic information and its application to the quasi-species model of biological evolution. BioSystems 2002, 66, 105–119.73. Denbigh, K.G.; Denbigh, J.S. Entropy in Relation to Complete Knowledge; Cambridge University Press: Cambridge, UK, 1987.74. Jaynes, E.T. The evolution of Carnot’s principle. In: Maximum-Entropy and Bayesian Methods in Science and Engineering; Erickson, G.J., Smith, C.R., Eds.;

Kluwer: Dordrecht, 1988, pp 267–282.75. Jaynes, E.T. Probability Theory: The Logic of Science; Cambridge University Press: Cambridge, UK, 2003.76. Wicken, J.S. Thermodynamics, evolution, and emergence: Ingredients for a new synthesis. In: Entropy, Information, and Evolution; Weber, B.H., Depew, D.J.,

Smith, J.D., Eds.; MIT Press: Cambridge, MA, 1988, pp 139–169.77. Adami, C.; Ofria, C.; Collier, T.C. Evolution of biological complexity. Proc Natl Acad Sci USA 2000, 97, 4463–4468.78. Atmanspacher, H.; Weidenmann, G. Some basic problems with complex systems. In: Large Scale Systems: Theory and Applications; Koussoulas, N.T.,

Groumpos, P., Eds.; Elsevier: Amsterdam, 1999, pp 1059–1066.79. Bar-Hillel, Y.; Carnap, R. Semantic content. Br J Philos Sci 1953, 4, 147–157.80. Standish, R.K. On complexity and emergence. Complexity 2001, 9, 1–6.81. Brandstatter, R.; Gwinner, E. Das Zeitgedachtnis der inneren Uhr. Max-Planck-Forschung 2001, 1, 60–64.82. Armitage, J.P. Behavioural responses of bacterial to light and oxygen. Arch Microbiol 1997, 168, 249–261.


83. Berg, H.C. The rotary motor of bacterial flagella. Annu Rev Biochem 2003, 72, 19–54.84. Kim, K.K.; Yokota, H.; Kim, S.-H. Four helical-bundle structure of the cytoplasmic domain of a serine chemotaxis receptor. Nature 1999, 400, 787–292.85. Szurmant, H.; Ordal, G.W. Diversity in chemotaxis mechanisms among the bacteria and archaea. Microbiol Mol Biol Rev 2004, 68, 301–319.86. Golden, S.S. Timekeeping in bacteria: the cyanobacterial circadian clock. Curr Opin Microbiol 2003, 6, 535–540.87. Stenmark, P.; Grunler, J.; Mattsson, J.; Sindelar, P.J.; Nordlund, P.; Berthold, D.A. A new member of the family of di-iron carboxylate proteins. Coq7 (clk-1),

a membrane-bound hydroxylase involved in ubiquinone biosynthesis. J Biol Chem 2001, 276, 33297–33300.88. Van Gelder, R.N.; Herzog, E.D.; Schwartz, W.J.; Taghert, P.H. Circadian rhythms: In the loop at last. Science 2003, 300, 1534–1535.89. Diamond, J.; Bellwood, P. Farmers and their languages. Science 2003, 300, 597–603.90. Jones, R.A. The Secret of the Totem: Religion and Society from McLennan to Freud; Columbia University Press: New York and Chichester, 2005.91. Durkheim, E. Les Formes elementaires de la vie religieuse: le systeme totemique en Australie. Alcan: Paris, 1912.92. Damasio, A. Looking for Spinoza: Joy, Sorrow, and the Feeling Brain; Harcourt Publ. Inc.: Orlando, FL, 2003.93. Roy, D.; Hsiao, K.-Y.; Mavridis, N. Imagery for a Conversational Robot. IEEE Trans Systems Man Cybernet 2004, 34, 1374–1383.94. Kripke, S. Wittgenstein on Rules and Private Language: An Elementary Exposition; Blackwell: Oxford, 1982.95. Slezak, P. A second look at David Bloor’s knowledge and social imagery. Philosoph Soc Sci 1994, 24, 336–361.96. Fairbanks, A. The First Philosophers of Greece; Paul, K., Ed.; Trench, Trubner: London, 1898.97. The Concise Oxford Dictionary of Quotations; Oxford University Press, 2003.98. Holt, J. Unstrung. The New Yorker 2006, October 2, 86–91.99. Sudo, Y.; Spudich, J.L. Three strategically placed hydrogen-bonding residues convert a proton pump into a sensory receptor. Proc Natl Acad Sci USA 2006,

103, 16129–16134.100. Gordeliy, V.I.; Labahn, J.; Moukhametzianov, R.; Efremov, R.; Granzin, J.; Schlesinger, R.; Buldt, G.; Savopol, T.; Scheidig, A.J.; Klare, J.P.; Engelhard, M.

Molecular basis of transmembrane signalling by sensory rhodopsin II-transducer complex. Nature 2002, 419, 484–487.101. Zhang, Z.L.; Huang, L.S.; Shulmeister, V.M.; Chi, Y.I.; Kim, K.K.; Hung, L.W.; Crofts, A.R.; Berry, E.A.; Kim, S.H. Electron transfer by domain movement in

cytochrome bc1. Nature 1998, 392, 677–684.102. Maclin, E.L.; Low, K.A.; Sable, J.J.; Fabiani, M.; Gratton, G. The event-related optical signal to electrical stimulation of the median nerve. NeuroImage 2004,

21, 1798–1804.103. Weatherford, J. Genghis Khan and the Making of the Modern World; Crown Publ., 2004.104. Carninci, P. Tagging mammalian transcription complexity. Trends in Genetics 2006, 22, 501–510.

SUPPLEMENTARY SOURCESFor many of the topics covered here

I can claim no expertise. The resources

of the WWW have proved invaluable in

providing some of the background ma-

terial. In particular, the following have

all proved to be valuable sources for

both history and philosophy.

The MacTutor History of Mathematics

archive (http://www-gap.dcs.st-and.ac.uk/

�history/index.html): created and main-

tained by John J. O’Connor and Edmund F.

Robertson of St. Andrews University, Scot-

land, has provided delightful reading on the

development of ideas in arithmetic, math-

ematics, and philosophy.

The History Guide (http://www.

historyguide.org/index.html): edited and

maintained by Steven Kreis, has supple-

mented by inadequate knowledge of

history.

The Stanford Encyclopedia of Philos-

ophy (http://plato.stanford.edu/): ed-

ited by Edward N. Zalta.

The History of Philosophy pages(http://www.friesian.com/history.htm).

The Internet History SourcebooksProject, maintained by Paul Halsall(http://www.fordham.edu/halsall/).

The Hanover Historical Text project,from which the quotation of Parmenideswas taken (http://history.hanover.edu/texts/presoc/parmends.html).

During the time over which this es-say has been developed, Wikipedia hasemerged as a useful and reliable refer-ence tool (http://en.wikipedia.org/wiki/Main_Page).

The quotations from Gibbon’s De-cline and Fall of the Roman Empire,another good read, are taken from theWeb version provided by Philip Atkin-son and maintained by the ChristianClassics Ethereal Library at CalvinCollege (http://www.ccel.org/g/gibbon/decline/home.html).

The web sites for NASA (http://www.nasa.gov/home/index.html; see particu-larly the environmental monitoring of

ocean and land at http://modis-ocean.

gsfc.nasa.gov/, and http://modis-land.gsfc.

nasa.gov/, respectively), the European

Space Agency (http://www.esa.int/esaCP/

index.html), and the National Oceanic

and Atmospheric Administration (http://

www.noaa.gov/), provide links to much

recent work on monitoring of energy

fluxes, net photosynthetic productivity

(NPP), weather, etc.

A compelling overview of the sea-

sonal flux of NPP can be obtained from

movies available at the Earth Observa-

tory site (http://earthobservatory.nasa.

gov/Newsroom/NPP/npp.html).

The Richard Dawkins essay “Viruses

of the Mind” is available on the web

(http://cscs.umich.edu/�crshalizi/

Dawkins/viruses-of-the-mind.html),

and the last chapter of The Selfish

Gene, in which the idea of memes was

introduced, is available at http://

www.rubinghscience.org/memetics/

dawkinsmemes.html.


Documents

Vehicles for Semantic Inheritance - University of … · The initial interpretational steps include protein synthe- ... velopments of abstraction in writing, algorithms, ... tional