BC-1992-Systemic Typology and Topology

7/29/2019 BC-1992-Systemic Typology and Topology

1/41

By way of illustrating the problem we are attempting to open up in this_paper, considerUre oUowing text. A is swimming at the beach and has ust spotted a black fin; B knows adolphin when sheseesone and tells A to calm down.A : H e l PB: - Don' . worry ,i E ' s j u s t a d o l P h i n .

A's mistake is hardly surprising. Common senre tells us to be worried about sharks atAustraiian beachesand suspicious of fins breaking the surface to compete with us for ournext wave. Scientificaily however the mistake seemsa little odd. How can it be that Acan mistake fwo animais that are so distantly related? After all, the biological taxonomyrelevant to sharksand dolphins looks is organisedas n Figure 1..1

javlessr sharkselasrnobranchiomophi.) L rags

etc.

-3 -15-

SYSTEMIC TYPOLOGYAND TOPOLOGY0. Taxonomyand ecologY

amphibiansvert ebrates

mamrnals

- J. R. Mart in 6 C Matth iessen

primates - whales, \ lcetacea t l L dolohinsr5batsetc.

Fig.1: Taxonomic separation of sharksand dolphins

Taxonomically speaking, dolphins and sharks are as different as monkeys and tadpoles.So it seemsodd that a body-surfer would achrally confuse the two.The problem is of course that scientific taxonomies are not based on differmces andsimilarities between animals that can be easily perceived. It's the genes and chromosomesthat count, not outward appearances. And how an animal looks on the outside is onlyindirectty related to its genetic make'up. Environment has a big role to play in how things

1 Latin ermshavebeenavoidedn this Euonomy herepossible.But here are no alternatives ssuoerordinatesor sharks nd ays,or for whales nddolphins.

+H:pu""nt"r'j


2/41

- 3 4 6look. Becausehey share n many respectshe sameecoiogical iche,sharksand rlolphinshave evolvedfrom quite unrelatedspecieso adapt o their environment n similar ways.They look sameand in many respectsunction n the sameway - but they are not cioselyrelated. For this reasonmost swimmersneedsome reassurance henever dolphins arearormd.The problem just iilustrated is a fundamental one for any form of taxonomicrepresentation. Taxonomiesare theoriesof similarify and difference with respect toparticular criteria. Once a criterion, or more usually a set of criteria, is adopted as aclassificatory principle then the parametersare set. Things are similar or different withresPect o thesecriteria - this is the information the taxonomyrepresents. But there arealways other criteria that could have been chosen, hat have been set aside. Andrelationships among the things being ciassified will necessarily ook rather differentwhenever thesecriteria are taken into account. So a taxonomy is only ever a partialstatementof similarify and difference.

Because f this it is very important in scienceo be dear about the parameters hat havebeen set and to have alternativeways of dealing with those that have been set aside.Some alternativesmay develop into researchparadigms n their own righq in biology forexample, he study of ecologyhasdeveloped o complement raditional taxonomicstudiesin this way and many universitiesnow havedepartmentsof environmental studies whichtake a more holistic and ideologicaily nformed approach o studying the world we arefailing to share with other phenomena.In this paper we are exploring whether or not acomparable environmental perspective s required in linguistic theory - especially withrespectto a theory like systemictheory which placesso much emphasis on systemicrelations.

1"Resource nd agnationin systemic heory languageasa whole is seenasa resource,organized nto a number oflevels or strata and diversified (most deariy at the content plane) into three generalizedmetafunctions.The interpretationof languageasresources arguably the centre n a clusterof reiated theoreticalassumptionsabout language: t leads o language unction in context,with text as the basicunit of language; t leads o a metafunctional subtheory' of how theresource s diversified, a stratal subtheoryof how the resources distributed into differentlevels of symbolic abstraction, an axial subtheory of the paradigmatic axis as thefundamentai organizing principle of each evel of the resource,and so on. The axialsubtheory s foregrounded n the nameof the theoryas a whole - systemic heory. It is, inthe first instance,a theory of choice and agnation. t is a theory of what options are

available to a language usr, how they are related (agnate),and how they are realized.The questionof options and their agnation s answeredby the system nefwork of systemictheory, which makes explicit how options form systems,setsof alternatives with enbryconditions,and how thesesystems,hrough their entry conditions, orm system networks.From one perspective, he system network constitutesa theory of fypology of linguisticunits. As such it differs in significant ways from other formalisms that might be used torePresnt such a theory - representations like the simple substitution sets used inpedagogicalgrammars,discrimination networks,and tables matrices)"Crucial differencesinclude the possibility of simultaneoussystems, ystemswith complex(coniunctiveand,/ordisjunctive) entry conditions, and recursive systems. A rypolory makes explicit howphenomenaare rdated; it is a way of interpretingagnation.But, as Lemke (to appear)hasPoind out (see urther below), we know from mathematicshat there s a complementaryperspectiveon agnation hat cuts asoss tlpology - topology.In this paper, we will explore the need for the fy?ologicalperspective n systemic heoryand a division of labour betweentypology and topolory (Sections3 and 4). But we willbegin by examining different kinds of 'multidimensionality' in systemic typology defined


3/41

by the different dimensions of systemictopology in favour of typology for certainpower of the r,vpologicalperspective is'

2.Agnation seen hrough the prism of systemicheoryBeforewe explorecasesof agnation hat are hard to representsystemicallyat present

then, et,sreview io* "gnatio.t".* be representedn current systemic heory' We will notethat agnation is ,.tol up?ur"ntedat one rank only nor at one ittatum only' Agnation' likeother lhurro*"rra, is dispersed hrough the prism of systemic heory'

2.1Basicagnationwithin system efworkBy taking a strandof systemsn a systemnetworksuchasthe strandof M@D systemsnthe clauseFammar of tnghsh, *e ian illustrate the basic orm of agnation in languageasrepresented n systernic t eory (note that the feature yes/no' could be added to thedisjunctive entry condition or ine TAGGING system o allow for the option of tagging ay.Jl"o interrogativeclause n AustralianEngiish):seeFigure2'

theory (Section 2) . That is, before rve adoptpurpo*s, we have to make explicit what the

'interrogativcges/no

Finite^Subject+ flood+FiniteMood( inite)[ +Subject lMood(Subj

tagqed+ floodteg+Tagfi 'ite+Tagsubjectf'loodtag^unt899e0

Fig.2: Agnation in one strand of an English grammatical system network

The agnation representedhere is internal to the syslem network: options such as'indiCative',imperative', declarative',and 'tagged'are related n terms Of the systemiCrelations of alternation within systemsand delicacy.The existenceof disjunctive entryconditions makes t possible o bring togethglalternativeswithin the system network toshowhow they shareoptions,as n ttie ca"sef TAGGING n the fragmentabove' (As notedabove, this is one of the respects n which a systemnetwork differs from the kind of stricttaxonomy that canbe encoded n a discrimination network)

The kinds of systemic oppositions rePresentedn this small network are illustratedbelow. Therealisationnrlei iin boxes)ottd"t featuresn the network show how the variouschoicesarenvnifested in stmcture'

vh-lfrh-llwh^Fin i teI

systen:Why don't yan try a topologtcalrioiation?CoAan't yw try a topological rientation?Yottcouli try a toql;gicA orimtation,kouldn't you?)Try a topological ientation.

Iwhllyes/nolIdeclarative]Iimperativel


4/41

''rv'hywhstrucfure:

metafunctionideationalinterpersonaltextual

textual:

d o n ' t y o uFirute Subject i - n t- - JPredicator

clausesystemsTRANSITIVITYM@DTHEME

a e o p o i o g i c a l c r i e n t a t i o n ?Complerrurt

22 SimultaneoussystemsStill within the system network, there may be one or more simultaneousstrands such asthe mood strand 1bove.The systemnefwork thus makes t possible o representmultipleperspectiveswhich intersectto yieid multidimensional paradigms. br moit general terms,theS multiple perspectives ypically correspondto the thrJe major metifunctions, aswithin the clause:

But they may aisoderive from within the sarnemetafunction;for exampie:ideational: AGENCY ('middle / effective')&PROCESS YPE 'material/ mental/ verbal/relational') &

various circumstantialsystemsMOOD TYPE ('indicative/ imperative') &POLARITY ('positive/ negative')&VOCATION ('vocative non-vocative')THEME ('unmarked/ marked')&TI{EME PREDICATION('predicated/ non-predicated') &CONIJIINCflON ('coniuncted/ non-coniuncted')

interpersonai:

Some of these simultaneous systemshave been interpreted as belonging to separatemetafunctionsin Fawcett's 1980)version of systemic heory. For instL.E, h" t"tt ,tpnegation as a sparatemetafunction. However, we can see theseadditional .o.t pon*ttsimply as more delicate distinctions of Halliday's four metafunctions (just as logical andexperiential are subtypes of the ideational metafunction). In fact, there aie strongarguments for not separating them out into different metafunctions. For instancelPOLARITY systems interact with MOOD systemswithin the grammatical system network(for example, in the absenceof negative exclamativeclairses),and. also within thesemantics.Furthermore, the realization of polarify is of the interpersonal mode - it hasthe potential of engendering negative prosodies (see Matthiessen, 1.988, or furtherdiscussion).The systemic ntegration of MOOD with POLARITY s illustrated in Figure 3.


5/41

-349

TEXTUALlauserammsr(THEHE CONJUNCTION)

i nterroqativeges no

Fini c'Subject+ flood+Finitetlood( inite)[ +Subject lf lood(Subj

positive

EXPERINTIALlauserammEr(TRANSITIYITY)Fig. 3: The simultaneous metafunctional perspectives

The structural consequence of system network internal agnation is that realizationstatementsall refer to the same structtrral unit - clause,SouP, or whatever.

2.3 Agnation outside the system nefwork: stratification and rankAs already mentioned, the possibility of simultaneous systems within the systemnetwork allows for multiple perspectivesand thus makes it possible to show agnation thatcannot be shown within a single strand of the system network. But agnation can also becaptured by going outside the system network of a given grammatical unit, either bymoving across ranks (within the same stratum) or by shunting between strata (acrossstrata). That is, we can show agnation in the sysEm network of another unit related to oneunder consideration via either the rank scale or stratification. When the representation ofagnation is dispersed in this way across more than one system network, the differentstatements of agnation are, of course, no longer related systemically; rather, they are

related ysaliz3tiinally, by means of preselection. Thus a semantic option rnay be realizedby the preselection (across strata) of a grammatical one at clause rank; and that optionmay in turn Uerealized by the preselection (across anks) of another grammatical option atgroup ranlc

v h -t-wh-llwh^r in i teIdecl rttive

taqqed+Hoodtag+Tagfinite+TagsubjectfloodtaguntE99e(


6/41

- 5 0 -In addition to agnarion within a single sffand of a systemnetwork, we thuspossibilities,viz. simultaneify within the same svstemnetwork, reiationshipnetwork at another rank, and reiationship to a systemnetwork at anotherdiagrammatically as in Figure 4.

have threeto systemstratum --

stratificaiion

Fig. 4: Dfferent dimensions for representingagnation

(The innermost circle is phonolo5y; it is endosed within lexicogrammar, which is, in furn,enclosed within semantics; for a discussion of this form of graphic representation ofstratification, see Matthiessen & Halliday, 1989.)There is one additional source ofmultiple agnation that we haven't mentioned or shown in the diagram -- grammaticalmetaphor (Halliday, 1.985: h. 10)"From the point of view of agnation, it is similar to theintroduction of a new stratum or a new metafunction" But it doesnot function as a new level;rather, grammatical metaphor is a way of using existing resourcesmore than once (to speaksomewhat loosely). Consequently, an example like aaious physical msuetneflts is relatedto other systems at dause rank in terms of transitivity and to other systems at group rankin terms of (among other things) premodification.Interpersonal grammatical metaphor has tended to develop as a theory of interactingtypologies. The various options in the discourse semanticsnetwork SPEECH FUNCTIONare related congruently to specific lexicogrammatical MOOD options (e.9" givinginformation with dedarative, demanding information with interrogative and so on).

Incongruent realisations (i.e. the indirect speech acts of speech act theory) are thentreated as special cases as grammatical metaphors (seeHalliday 1984,1985)" Reasoningalong these lines, a speech act like Why don't you try a topological perspectioe? ssemantically a demand for goods and services (a Command) realised incongruently as a whinterrogative (instead of the congruent imperative). This and some alternativepossibilities are exemplified below:

s'imultaneitg(metaf.)


7/41

a - 4- J J I -

Command= [demandinggoods& services]realisedbyWhy don't you try a topological ientation?Coutdn't you try a topological rimtation?You could try a topological riantation, ouldn't you?[Uust as a su1Sestion),ry a topological rientation']

incongruentincongruentincongnrentcongruent

One way of looking at this kind of interpretation s to read it as indicating that a speechact like 'Wy don't-you try a topologica[erspectiae? as seiected wice for M@D -being both iimperativel and lwh interrogative] in meaning; but one of thesemeanings istakei as'deeper, than the other. The dbpth metaphor predicts whi5h MOOD selectionwill be fypiiaily responded to - namely discourse semantics: Okay, we will beingpreferred'to Because e haont't got ime. It predicts as well the sequenceof resPonseswhen both selectionsare picked np - na*ely, lexicograrunar first (the 'surface'reading),then discoursesemantics econd the'real' reading)as n:Cou ldn ' t you t ry a Eopo log ica l pe rspect ive?- Y e s , a l r i g h t .

Frustrating the expectationsestablishedby theseprinciples is a source of verbal play.The second move in the following attestedexchange or example is best made whilesmiling (A is commentingon an interpretationby B of a piece of shrdent writing):A: (suggesting)why don,t you take a step or two further and talkabou! what it would. take to make the text subversive rather than- i r r a t - ^ h h a s i E i O n a l ?-JLr - u vl/vvB: - (smiling) r rike kids to be powerless '

Let,s consider an illuslration of the dispersal of the representation of agnationthroughout the linguistic system. Staying with our earlier MOOD example, we candesceid the rank r.it" from clause to group to caphrre distinctions arnong interrogativeitems along dimensions that are not syst^emicizldat clause rank (ruhoI what; what Iwhich;who"l whom). We can also move up and down the stratal organization.-Movingupwards to semantics,we can representagnation in terms of a speechfunctionai systemnetwork. gryecan aiso posit a higher-ranking semanticsystemnetwork of exchange,lvloseoptions are realized baspeech ftlnctional seiections,which, in turn, are realized by M@Dselections.)Moving do**".ds to phonology, we can representagnation in terms of anetwork of tone 6ptions (falling, rising, falling-rising, etc.). Diagrammatically as inFigure 5.


8/41

- 3 5 2 -

mood ofdeix'is

Fig. 5: Dspersal of interpersonal agnation

Note how tracking agnation across strata or ranks may be faciiitated by the notion ofmetafunction. Sinie ideational, interpersonal and texfual meaning is dispersed acrossranks and strata in the model, meaning at any point in the system can be relatedsystematically to agnatemeanings deriving from the samefunctional component. Pursuingthe interpersonal metafunction,ihe initiating move in the exchangebelow can be analysedacross strata as follows:

Why don ' t you t ry a topolog ica l or ientat ion?- Okay , we wi l l .

DISCOURSESEMANTICS ldemanding/goods&services/initiatinglLEXICOGRAMMAR [interrogative:wh/negative/addressee'subjectlPHONOLOGY ltone1]

And within the discourse semantics the following analyses of both the initiating andresponding moves can be provided (see Ventola, 7987, or the interpersonal analysesillustrated here):


9/41

- J J U -

IB-event/action]XCHANGE:

A 1

MOVES: [demanding/goods & services/initiatingl' ^ I l r r r r l n n r F r r n r t )r I I l J

A1Idemanding/goods&services/responding oJ- Okay we wi l l

Rwhen wor . re r s l - .uck.l . . v . . J v g Y v 9 v lt - o o e f a r o r r n c l i . t . .F - ; ^ JL ! I E q

Within lexicogrammar,Halliday 1985strongly foregroundsagnationacross anks withinthe logical subcomponent of the ideational metafunction by using the same strucfuralnotation for complexesat all ranks. For example, he clause,group and word complexeslisted below are analysed by Halliday as structurally identical:

clausecomplex:group complex:word complex group):Try Eopography

L L . Y

2.5Generalization across trataand ranksAs noted above, simultaneity of systems cross


10/41

are distributed (paradigmatically) across the various MOOD options in the clausegralrunar, as can be seen n the Table 1 below 6ased on Halliday,1957).

Table 1: M@D, KEY and TONEThe KEY systemsfor the'declarative' and'yes/no interrogative' environments are set outin Figure 5.

[- unmarkeddecrarativcJ m ;ungriiiedpolaritg I esserrionL ' r r t . a l l ' 3 1 l f E

L commiltedolaritg+ J"*.tionI EL contradictionr wh-Iinterrogative I J- unmerked' l I rT l- yes/noI l 'z I J- a=livcL marreaI l'4 I I Try

demrndedL non-"rr.rt iveI l ' r IL notdrmendod

Fig.5: KEY options realized by TONE selections(ii) With respect to semantics and lexicogrammar, the principle that the move down instratal organization embodies generalization may not be as clear as above withlexicogrammar and phonology since both semantics and lexicogrammar as part of the

polaritg of rnsvcrconscqucntielEpolritg of :nswcrnon-consrqucntirlE

KEY options crossmoodUpcs)

ooE

hdic.decl. possibi'litgstetement asscrtion rcscrvrtion contrrdiction,prolcstintrr. vh- (i ) ientrtive[Vh not tonic](ii) ccho

ycs no uninvolvcdqqcstion demldqucstion'involvcdqucstion rssertivequcstion

Posilivrimper':tivc*t"*

'inviirtion command 'ins'istcntcommand eompromisingcommlnd qucstionprohibition paremptorv: 3 [ - ] r l [ \ ] \ 5 [ ^ l ! 4 [.r] \J 2 t / l

TONEp i mar , no -comPound


11/41

'content plane' (in Hjelmslev's sense)and they are related 'naturally'. However, if weconsider the idea that different contexts of situation correlate with different sihration-specific variants of the semantic system - such as the regulatory semantics discussed inAalliday Q97, -- the difference in generality befween semantics (situation-specificsystems) and grammar (one generaiized system) can be seen very clearly. In other words,tLe co.rtet t system of language has to cope with the tension between unify and diversity;and it does this by engendering diverse situation-specific semantic systems and onegeneralized grammatical system as shown in Figure 5.

context (different coniextualsettings)functional reg'ister) variation

Fig.5: Diversity and unity within the contentplane'In the examplewith KEY and TONE, the environmentsacrosswhich the diverse options

in IGY are distributed are internal to the system the options defined by MOOD. ln thepresentexample, he environmentsare external o the semantics the different contextsofsituation.Rank. Since he linguistic resourceswithin a given stratumare ranked, agnationscan bestatedmultiply at different ranks and, as the rank scale s descended, he agnationsmaybe generaliz-edacrossdifferent higher-ranking contexts.Consider the distinction in thetrer6al group between perfective and imperfective aspect, .e., between infinitival andparticipial form. Looked at from the vantagepoint of word oI gEguPluttk, there is thisiingte systemicdistinction: 'perfective/ imperfective' e.g., o do doing . However, if wetate ttre clause as the vantage point, we find the more functional differentiation; the

single verbal group system 'perfective / imperfective'corresPonds o a number offunctionally distinct systems n the clause,each of which has a pair of features, one ofwhich preielects 'perfective' while the other preselects imperfective'. As a result, therealizational movs down the rank scale s, in this case,a generalizationof a number ofdistinctions - diagrammaticallY:

lex'icogrammar

phono'logg


12/41

clause

grouP

Fig. 7zDifferentiation (clause)and generalization (group)For instance, he following ciause ypes have different'aspect' systems,corresponding tothe singie verbal group system perfective / imperfective:

- rankshifted, elaborating dause lrestrictive relative dause]: moduiated /unmodulated(modulated:)t fre man Eo do the job is Henry'whocan/shoulddo'(unmodulated:)he man doing Ehe lob is Henry 'who does/isdoing'

- dependent,enhancingclause: rrealis / realis(irreal is:)Henry left Eo rehearse his part(real is:)Henry lefE rehearsing his part

Here are some further examplesof ciause contrastsother than the two mentioned above:- dependent, extending clause:opposition grammatically conditioned by aconnective such as rather than and instud :

Rather than leave, Henry rehearsed in the kiLchenIns tead o f leav ing, Henry rehearsed in the k iEchen

- rankshifted, 'act' dause as serving as Phenomenon in perceptive clause:bounded / unbourded:(bounded:) We saw Henry leave when we arrived(unbounded:)we saw Henry leavinq when we arrived

- rankshifted,'act'

clause as Subiect:To be or not to be is the cruesEionBeing aE home al l day is not the problem


13/41

A similar kind of examplescan be drawn from MODALITY in English. Halliday Q970;1985) shows how there are fwo broadly different systems in the ciause -- indicativeMODALITY (MODALZATION), the scalesof probabilify and usuaiify from possible and'sometimes' to certain and 'always', and imperative MODALITY (MODULATION), thescalesof obligation and inclination from aliowed and willing to required and determined.At verbal group rank, these fwo realms of modalities are generalized as the system ofmodal auxiliaries (cf. Halliday,7970): seeFigure 8.

c lause :

group, erbal :

Fig. 8: Generalizationof MODAI.IZATION and MODULATIONin system of auxiliariesThe difference across clause rank and group rank has considerable significance for thestratal relationship between grarrunar and semantics. In a systemic account, therelationship is stated in the first instance at the highest rank at which it occurs and sincea systemic grauunar of Engiish is clause-based, he relationship will be stated for clausalsystems, where the agnations are distributed acrossdifferent (paradigmatic) ccntexts suchas enhancing vs. elaborating. In contrast, traditional grammar started with lower-rankingsystems; it was essentially word-based. So we would have to try to come up with ana-bstract pair of semantic glosses for the group distinction perfective / imperfectivereflecting the degree of generalization the distinction embodies and we would also have tolist the various uses of the forms out of their clausal contexts. Similar examples could be

glven from TENSE, VOICE, TRANSITIVITY, PERSON,and so on.2.5 Dispersal acrossstrata and ranksWe have noted that moves down across strata and ranks often entail a generalization:more options (in different environmenb) are realized by fewer options. One case n point is

modalization:required- determinedsupposed- keenalloved - villing

modalization:certain-- alvagsprobable - usuallgpossible - sometimes

sgstem f modalauxi]iaries:mustwillmav


14/41

MODALITY, rvhere !{ODALIZATION and IvIODULAION are generalized n the systemof modal auxilianes. \,ICDALITY aiso illustrates another phenomenon in the move fromhigher to iorver sffata/ ranks. This is the dispersal of realizations of a highgr sy9!1macioss different environments at a lower stratum/ rank. Thus the system of MODALITY(MODAL17AION and IvIODULATION) at clause ank is realized at group rank3(i) adverbiai ly (MODALIZATION): perhaps,possibly, certainly; sometimes,usually, always &c.(ii) verbally -

(1) through modal auxiliaries (MODALIZATION/ MODULATION): zay,will, must; &.c.(2) through conative verbal group cornplexes MODIJLATION): is obligdto, is supposed o, is allowed o; is detnmined to, is kem to, is willing to& c .

The higher system is more holistic; it spans the full range of meanings realized by thelower syste*s. Consequently, it is possible to show agnations between sets of items that areconstrued in differeni environments by the grammar, e.g. perhaps probably cntainly ::may : will : must :: possible probable certain. Grammatical metaphor may createagiations between environments that are dispersed too far aPart within thelelicogrammar for them to be brought together easily within a more holistic gtuTTutij]lryrt"ri. For instance, grarunatical metaphor createsan environment for MODALIZATION'outside' the clause iit ttr. congruent- reading) -- in clause complexes such as I thinklbetimel guess I I theyhavealready eft (see Section 3 for further discussion).

2.7 Plane: the contextsof languageBefore turning to the question of missing agnation in section 3, two further descriptiveresources need- o be considered. The first has to do with the semiotic environment inwhich language means - the issue of planes (this section). he second has to do with one'sorientation to bottr system and text as process and product - the issue of synoptic ordynamic perspectives (Section 2.8 below).Pursuing the first issue here, the point that needs to be made is that the notions axis,metafunction, rank and stratum developed to this point have to be contextuaiised.I-anguage realises context - it constmes, is consEued by and reconstrues, symbolically, thesociit slstem (see Halliday 797$. This means that systemic relations at all levels within

language can be projected onto a model of context in such a way as to bring out furtherdiriens-ions of agnation. The contextual systems are interpreted as connotative semioticsystems, i.e. seriiotic systems realized by other semiotic systems (cf. Hjelmslev's 1943discussion of denotationssprog and konnotationssprog). Martin 1985 provides oneextravagant example of a projection of rhis kind, with three orders of contexhral semioticsystems- ideolory, gerue and register: seeFigure 9.

3 ilAOORUITmayalsobe realizedn ye totherenvironments(i ) think...perhaps'&c.;and(ii) t 'swssible that .. it is a kindof systemic rosody; nd as such t increaseshe potential or prosodiciealization,e.g. t think it may perhapsnot be a bad deaafterall.


15/41

phonologg

Fig.9: I-anguageand its semioticenrrironmentSetting aside the levels of ideology and genre here, the register variable tenor for""u*plE can be used to further develop the interpretation of the exchange Why don't yoytry a'topological rientation? Okoy,wewill. Following Poynton 1984,.1985 e can askquestions lbout the kind of interpersonal relations realised: is the status of theinterlocutors equal or unequal as reflected in the reciprocity or non-reciprocify of thechoices selected?;are the interlocutors in frequent or infrequent contact with eachother asindicated by the proliferation of choices aken up and the degree of contraction in theirrealisation; and what degree of affect is realised, measured with respect to amplification.The example being considered s too short to provide satisfactoryanswers to questions sucf

as these. But once considered as part of a longer text, the exchangecould be interpretedalong these ines - recontextualisedas it were on a higher plane.Tenorvariable:STATUSCONTACTAITFECT

realisationprinciple: tentativeanalysis:(reciprocity) unequal?(proliferation; contraction) low?(amplification)

28 Perspective:synoptic and dynamicMartin 1985 explores the question of simultaneous synoPtic and dynamic perspectives onsystem/text, p


16/41

comings of this stancecan be exemplified with respect o exchange trucfure by consideringtextssuch as the foilowing (symbolsas n Ventola 1987):

Why don't you try a topological perspective?- Topological?- Yeah, you know, like in maths.- Huh?- You know that paper of ]ay's -- No.. .

A2cfrcf K1- c f- rcf K1K2fHere, the initial demand for goods & servicescannot be negotiated until its experientialmeaning is further explored. But this exploration is unsuccessful,since the interiocutorbeing positioned as secondary actor has never heard of topoiogy, does not appreciate itssignificance in mathematics and hasn't seenJay Lemke'spaper. Not only does the initialexchange not reach closure, but the interlocutors end up negotiating something verydifferent from what they began with. Interactions of this kind are quite systematic, butnot predicted by system/structure theory at the rank of exchange;nor, Martin argues, s itobvious how present qrnoptic descriptions could be extended to account for this and relatedphenomena (seeVentola 7987 for further discussion at the level of genre). The generalthrust of the argument is that by emphasizing the paradigmatic, and abstracting awayfrom direct representation of sequence n text, systemicistsput themselves in the position ofnot being able to account for choiceswhich depend on just where in the unfolding of a textthe realisation process has reached (unlike Firth in other words, elements of sfructure donot act as entry conditions for paradigmatic selections). Accounting for these strucfuredependent choices, which take into account the meanings that have so far accumulated

and where the text is going next then becomes the responsibil ify of dynamicrepresentations, a frontier area of research n systemic theory (see Mathiessen 1988;Bateman 1989).

2.9 SummaryAl l of the various dimensions discussed to this point in the paper are presentedrypologically in Figure 10 (a network at the level of field, mapping part of theorganisation of systemic linguistic theory)" This ailows us to position the new dimension

being opened up in this paper - the typological/topological one - as one with a synopticand a paradigmatic focus (thus the conjoint enbrycondition).


17/41

- i D i -

' ideational'interpersonaltextualconnotative

denotative

sgntagmaticparadigmatic

discourse F clausesemantics I; ,va leu r , { n .ouo /phraseI LI worolexico-grammar IIL 'co l loca t ion '

phon/graph-ologg

\) 1 tgpolog'ical- -L toPoloqical

1-srJnopt'ic-)L dgnamic

Fig. 10:Metalinguistic resourcesTaking into account the dialectic of realisation interfacing these dimensions, analternalive form of representation, in the form of a helix is presented below- Thispresentation is topological rather than typological - a spatial metaphor is. used to connect

ievels as a procesi of ieatisation. For reisons of spacewe cannot explore the helix in detailhere. Basically it maps realisation as a two-way process looping through strata andplanes while t-eturning at each level to recurrent metafunctional themes (the 'columns'ihat can be read "qfosl levels). It is to topological representations of this general kindthat the paPer now turns.


18/41

- J U - -

race gender class

culminative

3. Agnations not accotrnted orWe have seen how agnation may be'dispersed' across metafunctions, ranks, strata andplanes and that complementary synoptic and dynamic perspectives can be adopted todevelop further inter-connections.This makes it possible to capture relatednesswithvariation in metafunctional domain as well generality. In spite of the various possibilitiesfor representing agnation that we have reviewed - all of which are, of course, quite well-

codingorientat. ion/ hetcroglossia

pariiculate


19/41

known - we encounter agnations that are hard to rePresent n a satisfactory or clear-cutway. We wiil draw examples rom all three metafunctions.

3.1An experiential examp e: behavioural processesAn example from experiential ciausegrartmar will be considered first. Halliday's 1985account of p.o..s type distinguishes three main classesof clause - material, mental andrelational; iUua"y'goes on to consider three additional tFpes*hi+ are-ciosely relatedto each of these - behavioural, verbal and existential respectively. If we formuiate theseproposals systemically, adding labels to generalise across material & behaviouraiiaoittg), mental & verbal (semiosis)and relational & eistential (being), we arrive as aprocessqrPenetwork such asthe one n Figure 11.

doingmoterisl

behavioural

They attacked

Tb.ey frowned

;- verbalIsemiosisI ; PerceltiveL mentel+ alJecti-veL cogrrr-itive

Tbe eskedThe sew it

rel etional

Tbey liked itTb.ey knew itTb.ey were foolisb.

beingex'istetial Tb.ere were tanJ


20/41

What seems o be going on here is that behavioural processes onsfruct saying, thinkingfeeiing and perceiving as action, in terms of their outward physical manifestations. Assuch tJreyare all associatedwith one or another verbal and mental process,and in somecaseswiih relational processesas well (becauseof an association between mental andrelationai processswhich will not be further explored here). Some examplesof theagnation series suggested here are listed below (for further discussion of theserelations seeMatthiessen 1989):

ta lk / sayHe is ta lk ing to

Ibehavi-oural il { o q r r , r s L h a c h e ' sEhem r ighE now.

r a E h e r c o n c e r n e d .' l i < f o n / h a a r

T t m l ' i s t - e n i n o a E . E h e m o m e n t .f hear that you 've been concerned for some Eime

s m i l e / l i k e / b e h a P P Yr r ^ . : ^ ^ - { 1 . ; - - a E E h e m .1 e J - > > r t M r r r vH e l i k e s t h e m .

I m e n E a l : r e a c t ]He is happy wi th them.

ponder,/wonder/be in doubtHe is ponder i -ng Lhe i r ques t ion .

Ibehavioural ]He wonders whether Ehey were r igh t .He is in some doubt about Ehe i r concerns .

Iverba l ]

f h o l - r = r r i a r r r a I lL v v a l s v + v s r s 4 lI m e n t a l : p e r c e p ]

f l - r o l r : r r i a t r r a I I

f r o l : i - i n n : l IL 4 e 5 $ e 4 v . . $ 4 J

ImenEal : cogn i t ]l r o l a t i n n a ' l IL ! v - s v

The process ffpe network outlined above does not capture these similarities. It claimsthat behayiourat and mental processs are quite different (like sharks and dolphins). Itwould of course be possibtL to revise the network, moving behaviourals into thesignification class. In effect this would be to glve more weight to the presence of aco'nsciousparticipant in behavioural sand their ability to project Oy quoting): see Figure72.

metei al They rre atteckirr,gl nowbehoviouralverbalmental

Tbey ere talkingrlhey say they can'tTbey feel tbeY shouldn'tThey ere b,rPPY beY won't

typology Gi)

semiosi !

Fig.12 Process


21/41

l a :J U J -

But note that this would still no t capture the relationshipbetween behavioural andcertain attributive relational processes. And more seriously it loses the previouslyestablished elationship between material and behaviouralProcesses s far as tenseandreporting are concerned:TENSE:

PROJECTION:*Thcy talked that theYcouldn't.Thcy said they wouldn't.

What we find here in other words is a genuine tension in the system. There areparameters aiong which behaviourals esemblematerial Processes nd other parametersAong which they resemblemental and relational ones. Depending on-which parametersure tiketr as criterial at primary delicacy, behaviourals can be grouped with materials orwith mentals and verbais but not both. The criteria deployed at primary delicacy thusbecome thematic for the description. They act as its point of deparrure and circumscribethe agnation which can be established. In typological work, all points of departureforegiound some criteria and backgroundothers. With phenomenu l:h as behaviouralpro&r*r this means that some agnltion must be lost in the processof classification. Thetypology is limited in just this respect.

33 An interpersonal example:MODALITYLet us now pursue this problem with respect o the interpersonalmetafunction, focussingon MODAL1IY. Centrai to any discussionof probability in English the considerationofmodal verbs and adjuncts (Halliday 1970/1976;7982;1985).Median probability fo rexample can be realised verbally as Finite or adverbially asMood Adjunct in the clause(or'prosodically' in both functions):

modalised FiniteMoodAdiunctBut beyond this, there is a range of M@D resources that can be used to mediate thecertainty of a proposition, including tagging and interrogatives:

IE ' s behavioural, len' t l t,? declarative aggedIen't l t behavioural ? negativees/noWby wouldn' t it be behavioural ? negativewh modalisedAnd moving acrossmetafunctions, experientialoptions can alsobe brought to bear on thequestion of speakercertaintY:

mental projectionrelational fact

They are talking about it.They say they will'

I t wou ld be behav ioura l .P r o b a b l y i E ' s b e h a v i o u r a l .

f r e c k o n i E ' s b e h a v i o u r a l .I ' D s u r i t ' s b e h a v i o u r a l .

behavioural: present n presentmental: present

behavioural: no hypotacticprojectionmental: hypotactic projection

As with SpEECH FUNCTION and MOOD, this kind of dispersal of agnation has beenhandled through the concept of grammatical metaphor (e.g. Halliday 1985:33+340)-Developing an i-nterpretation along these lines would mean establishing a generalisedprobabiuty"network in the discourse semanticsand allowing for diversified realisations nlexicogrammar.


22/41

DISCOURSE EMANTICS:median degreeof certaintyAlternative realisations:I T r o a i < a n i l - t q h o l . r a r r i a r r r r " l IL - Y r v q ! { 4 . II T ' m q r l r o i l - ' q h o h a r r i n r r r a ' l l l v q r q ! . , lIL wou ld be behav ioura l .P r o b a b l y i t ' s b e h a v i o u r a l .I t ' s b e h a v i o u r a l , i s n ' t i t ?I s n ' t i t b e h a v i o u r a l ?'v {hy wouldn ' t i t be behav ioura l?

mental projectionrelational factmodalised FiniteMood Adiunctdeclarative taggednegative yes/nonegative wh modalisedOne of the main weaknesseswith this approach to mediating certainty is that it doesnotspecify how the alternative realisations are graded with respect to each other. It sayshow they are the s:une semanticaliy (i.e. triggered by the same discourse semanticfeature), and it shows how they are different grammatically. But it does not say howthey are semantically different. Simply augmenting one typology with another does notcapture the various ways in which alternative realisations complement each other in

occupying the same semantic space. For this some mechanism is required which will showhow reaiisations are more and lessalike, not just how they are typologicalty different.

3.3 A textual example: thematic resourcesBefore pursuing this problem of gradiencebelow, let us consider one final example fromthe textual metafunction. Hallidav 1968,1985(seealso Chapter 12 in Kress I97O hasnoted the significance of first position in the English clauses or the realisation of Themeand considered the way in which different selections for Theme may be more or lessmarked for a given MOOD. In declaratives or example,a Complement/Theme is marked

in comparison with Subiect/Theme.

Developing this point, Halliday notes the significanceof IDENTIFICAION -- anexperientialsystem put to work by the textual function with a view to organising Themesin reversiblerelationalclauses:

( \ z p o l o g y ' s a w o r r y ) , b u t t o p o l o g v I , d r e c o m m e n d .Topo logy 'd be recommended by Jay .

Topologly is what Jay would recommend.What Jay would recommend is Eopologry.

markedunmarked

THEMEIDENTIFICATIONTaking grammatical metaphors of the ideational variety into account, nominalisation isa further resource that is used to group meanings as Themes:

Jay's reconrmendation would be topology. nominal isat ionOnce again, there s nothing in the $ammar that showshow thesedivergent resourcesormarking Theme are related to eachother" And resorting to grammatical metaphor andstratification to handle this area once again fails to account for gradience. The point isthat quite different Parts of the gmrrnur as it is currently systemicisedcan be seen incertaincontexts to do very similar work. Typolory does not account for how this work isparcelledout acrossgrammaticalsystems.


23/41

3.4An interpersonal xample MODALIfi; gradience sosssystemsBefore urning to the problem of formuiating topologrcaldescriptions n Section4 below,let us return onceagain to the problem of gradienceor dines. To this point we havebeenmainiy concernedwith gradienceacrosssystems - how it is that distantly relatedgrammaticalclasses onstructclosely elatedmeaningstBut gradiences alsofound within

larticular areasof the grarnnurr,amongdosely relatedsystems.The areaof MODALITY,introducedabove,providesoneclearexample.As noted by Halliday 1985,MODALITY and MODULATION canbe interpreted in termsof degreesof certainty and obligation associatedwith propositions and propg*l?respeJtively. His interpretation of thesescales, ncorporating the closely relatedsemanticsaieasof usuaiity and inclination s outlinedbelow in Figure13.

-

supposed

allowed

MODULATION'imperative'tgpe

d o !

MODALIZATION'indicetivc' tgpc

it islprobablitV

ccrtainlg

probablg

possiblg

[usualitg

alvags

lobligaiion

required

usuallg

somct'imes

is sn't

An interpretation such asrepresented ystemicallYasa

don' tnegative

Fig.13:MODAUTY ascline

this suggests hat a system like MODALITY shouidcline:seeFigure14.be

hishmedianlov

Fini e: flodoladjunct:must certairr.lyF:i11 probablym8y gossibly

it must be

it v'ill bc

it mag br

must do

vi l l do

mag do

VALUE

Fig. 14: MODAUTY values(i)


24/41

- 3 6 8 -

But there are shong grammahcal arguments for not representingMODALITY gradientlyin this rvav. This is that median modalities interact with negate-dhesesdiffeiently thathigh and low valued ones. Consider the following proportion"alities:I t i s p r o b a b l e t h a t i L d i C n , t w o r k c u E , :i t i s no t p robab le tha t i t worked ou t :f L w o n ' t h a v e w o r k e d o u L : :I t : - s p o e e l b l e t h a t i t d i d n , t w o r k o u t :I E ' s n o t c e r t a l n t h a t i t w o r k e d o u t :I t may not have worked out : :I t i s c e r t a l n t h a t i t d i d n ' t w o r k o u t :I t : - s n o t p o s g i b l e t h a t i C w o r k e d o u t :I t c a n ' t h a v e w o r k e d o u t

The problem is that If 's possiblehat it didn't means he samething as If 's not certainthat it did at the same ime as lt's certain hat it didn't means th i same as It's notpossiblehat it did; but with the median modality, It's notprobablehat it rtid means thesameas If 's probablehat t didn't. This line of argumentationeads Hailiday to opposemedian modalities to outer ones n the grammar,network as below. In this nerwoik thetermsare not graded,and so their complementarityn thesemantic paceof certainty s notadequately mapped.

rneol8Fini e: Modal djunct:g.i11 probablymust certairr-lymay possj-51y

h i g houter

Fig. 15:MODALITY values ii)

3.5Over-detersdnedborder .ueas: lendsWe have seen that there may be border areas n a semantico-grammaticalspacesuch asthe spaceof processes; or instance,behavioural process ie berweenmaterial ones on theone hand and verbal and mental oneson the other. However, there is also another fype ofborder phenomeno. : the problemof grammaticalblends. Here we face he probiem ofooq-detumination. Particular phenomenaare equally well classes n differeni ways; thetypology does not resourcesor distinguishingbetrveenhem.One well-known example of a blend of this kind has to do with drawing a line betweencertain receptivemental Processes nd attributive relational ones. I'm sirprisedhe'sherefor examplecan be equally well treatedasmental or relational:

HODAL TYVALUE


25/41

- 3 6 9 -

IVIENTAL RELATIONAL (overdetermination):I ' m s u r p r i s e d h e ' s h e r e .I am surPr isedSenser Processl a m

Carrier Process

t b y t h e f a c t t h a E ) h e ' s h e r ePhenomenonsurpr ised (about t .he facE thaE) he 'shereAtlribute

Note that it would be misleading to treat m surprisedhe's here as both mentai andrelational as resorting to the notion of grammaticaimetaphor might imply._Blends are notmetaphors - one thing on the'surface'and somethingelse'underneath'. Rather they arehybrids - partly one thing and partly another. Typologically they are overdetermined --and stratification does not sort the problem out. There is a genuine gradience Presentwhich a typotogical orientation obscures.Another example of blending is found in the areaof symbolization (seeHalliday 1985:154). Verbal, mental and relationai clausesof the following kind are so closely relatedthat it is difficuit to find criteria allocating them to one Process ype rather than theother.4 And once an allocation has been made, them something of the sirnilarity amongthem has been lost.

VERBAL:The resul tsSayerMENTAL:

The resu l tsAssigner

ind ica tehocess

proveProcess

to meReceiver

Senser

Eo meAngle

The results convinceLnducer ProcessRELATIONAL:

Ehat t ypo log rY won 'E work .(Verbiage)

(o f Ehe fac t ) EhaE EYPoIogYw o n ' E w o r k .Phmomenon

( the fac t ) EhaE tYPologrYw o n ' E w o r k .Token

4. Topological descriPtionUp to this point in the paper we have focussed on Problems - the lirnitations oftyp'otogcaf aescription in tle-face of an accumulation of linguistic phenomena- We willrio'* t"* to the problem of addressing thesephenomena in positive terms, developingtopology as a complementary orientation with strategies for handling agnation whichor"retfiyrtemic description rlsources do not nranageeffectively. Note nlt-* taking thisstep we are arguing that the phenomena we .ue considering are unified and can bepoiiUvely ctraraiterised in somi respect they are not simply a rag-bag of left-overs that

tlTological analysis fails to explain.4 Note hat heverbalexamples verycloseo therelationalprove'example;f, however,heSayersa speakerorwriter),heclauses removedrom hisareaof blending: arlandCharle_sndicateo metnit typotogywont'work 'typotqy wont work'. lt is not.longerossibleo interpretCarland Charlesas beihg etiteOasTokeno that ypoloWwontworkas Value.


26/41

- 370 -4.1The notion of topologyLemke (no date), in a very helpfui unpublished paper, introduces a complementarytopological perspective at the level of genre, making use of arguments closely related tothose developed in Section 3 here. He characterises is topological orientation as follows:

A topology,nmathematrcalterms,sA SETCF CRITERIAORESTABLISHINGEGREESFNEARNESS R PROXIMITY MONGTHE MEMBERS F SOMECATEGORY.t turns'collection'or set of objectsntoa spacedefined y therelationsf thoseobjects.Objectswhichare morealikeby the crrteria re epressntedn hisspaceas being ioser ogether;hosewhichare less al ikeare further part.Therecan be multiple ri teria, hichmay be moreor lessindependenti one another, o that wo tens, for instance, ay be closer ogethern onedimensionsayhorizontalistance), ut urther part n another vertical istance).What sessential, bviously,s our choiceoi the criteria,he paramelers,hat define similarity nddifference n eachdimension. heseparametersanbe representeds moreor lessalike.Thesameset of parameters llowsus to describe oth he similarrtiesnd he differences mongt6ns, or texl-typesgenres).Lemke, .d.]

We can explore the reiationship betwem f.vpology and topology by means of twoanalogies (see Figure below). Imagine a system of rivers originating from one spring (1),flowing happily in separatebeds and branching off (2) into additional rivers, fanning outaiong a coast line" The coast line is very flat and the soii is soft so the estuaries of whatwere once separate rivers will merge together. Similarly, imagine frying to represent thedifferences and similarities between the darker areas of the figure to the right below. Ifwe take each black core as the starting point, the grouping is clear enough; but theperipheral areas are much more difficult to assign o one node or another in the taxonomydesigned to show how the dark areas are related. The are dearly grey border areas.

Fig.15:Typologyandproximityln neither casewould it be appropriate to abandon the distinctions made. The earlierstages 1and 2) are quite distinct and while further differentiationsmay turn out to mergewith others they need not. And, in the figure to the right, as ong as we focus on the core,the darker areasare also quite distinct. Sowe needsome orm of interpretation that will

allow us to capture the higher-level and coredistinctions as well as the grey boundaryareas"Both of the examples given above concern arangements in physical space;typological considerations oncerninggrouping according o differenceand similarity are amatter of relative location in that space. Expressed sa proportionality:similarity : difference ::proximity: remoteness


27/41

Now, it is possible o transfer he reasoningaboutsPace o abstractdomains, as we dowhen we talk of semanticspaces or, for that matter, semantic ields or domains). Indeed,this is not a new perspective in systemic theory. Halliday (1951:Section 2.2) identifiescline as one of three lrrpesof scale:... I haveused he erms'hierarchy',taxonomy',and cline'as general cale ypes.A hierarchyis aken o man systsm f erms elated long single imensionhichmust eone nvolvingsome ormof logical recedencesuch s nclusion).taxonomys akeno mean specialypeoi hierarchy,n-e itir woadditionalharacteristics:i) heres a constantelation f each ermto the erm mmediatelyollowingt,anda constanteciprocalelationf each o hat mmediatelypracedingt; and (ii)degrees significant,o hat he placen orderoi eachoneof the erms,bt"tables ine distince n number l steps romeither nd, s a defining haracteristicf thatterm.A cl ine esembles hierarchyn that t involveselationlonga single imension;utinstead f beingmadeupof a number f discreteerms clinesa continuumarrying otentiallyiniinite radation.

We can thus seea 'space'as being extendedaiong one or more clines.SA cline or a set ofintersecting clines may havea 'core'region - as n the traditional notion of 'core meaning'or 'cardinal vowel'. Movement away from the core along any of the clines will to moreperipheral regions - more peripheral members or subtypes, or, we.could say, moredistintly relited', in reference to Wittgenstein's family metaphor (cf. Daniel Jones'conception of the phoneme as a family of sounds).This phenomenon has studiedextensively in terms of the notion of prototype in a seriesof studies by Eleonor Rosch andothers; it iras been applied to the nominal and verbal domains in grammar by Hopper &Thompson 1983. Protbtype' is perhaps ypically characterizedn cognitive terms; but thisis, of io.trse, in no way necessary. t is perfectly possible to reason about ProtolyPessemiotically.Before further developing the interpretationof agnationand 'space',we'll introduce somelinguistic examples.The best known are probably phonologrcalor lexical. For instance,voiel systems re interpreted in terms of a vowel sPacewith cardinal vowel locations;

allophonic variation oicrrrs in the regions around the cardinal locations. Thus animaginary four-vowel systemcan be describedboth tFpologcally, as front/ back & high/low, and'topologically; the diagram in Figure 15 illustrates the complementarity betweenthe fwo perspectives. Ifront back

h igh

Fig.16:Vowel typolory and opology5 Sinc" Halliday's 961discussionf clines, imilar bstractionsavebeendiscussed utside lsystemicinguisticsn approacheso syntactic ategories squishes,ontinua, tc.

O . ]i6B


28/41

Here however we'll draw on the grarunar for examples.

42 Processes s topological spaceWe will pursue the notion of spaceand parameterswith respect to experiential clause

grarunar (see Section 3.2 above), elaborating on the kind of material, behavioural,relational tension introduced above. To begin we will circumscribe the semantic spaceunder consideration, with four processtypes defining points on its circumference: material,verbal, mental and relational. This gives us the frame of reference for our topologicaldexription - the semantic space bounded by notions of happening & doing saying sensingand being & having.Having located the focus of our description in this way we go on to define topologicalparameters. Proceeding in this way these hrrn out to be characterisations of the gradiencewe can specify among the four points defining our serrnntic space:seeFigure 17.

happeni g&,do ing sagingmatarial verba l

Iexter olse sing

IsUmbolizctionas dentitUli nternalsagi gI

relational mentalIsensings

,oa.uasmotion-behsvioura' l : 'sagi g'\ as ctivitg

bchavioural :sensing soctiv i tg,/

sgmboliz.scommunication

bcing &hrvi ngsensl 90tatt i bute i ncrtprocess scnsing

Eig. t7 zTopological parametersor English TRANSITIVITY


29/41

- -

Topologicalparameters

The clinesare exemplified below.oTypologicaloppositions(1) behaviourai region(i) behavioural: material-verbal:" C o m e h e r e " , h e g r o a n e d ." C o m e h e r e " , h e a s k e d .(ii) behavioural: material-mental:

saFng as activifysayrngascommunication

He watched them pulling Ehe wa1l down. sensingas activityHe saw Ehem pulling the walL down. sensingasinertprocessing(2) Spatial region: material-relational:The Eroops surrounded Ehe embassy. sPace smotionTrees surrounded Ehe embassy. spaceas relation(3) Symbolization region: verbal-relational:The report indicaEes they're inEell igent. symbolization/communicationThe report ref lects their inEell iqence. symbolization/identi ty(4) Semiosis region: verbal-mental:He convinced them they'd won. externalsensingHe concluded Ehey'd won. internal saying(5) Affection region: mental- relational:He d like them Eo come sensingas nert ProcessHe'd be pleased for Ehem to come. sensingas attribute(5) Qualitative change: material-relational:His grrip weakened changeas eventHis grip became weaker changeascomingtobe(7) Representation: material-relational:He is standing for parliament representation as activityHe stands for honesty rePresentationas (inert)svmbolization

6 To avoidclutter n the igurewe have eftout someclines.For nstance,here s, arguably, clinebetween eceptivepassive)material nd ntensive scriptiveelational nd a similarone betweenreceptivepassive) tfectivemental nd,again,ntensive scriptiveelational.


30/41

- i -

Topologicalparameters

The clinesare exemplified below.oTypologicaloppositions(1) behaviourai region(i) behavioural: material-verbal:" C o m e h e r e " , h e g r o a n e d ." C o m e h e r e " , h e a s k e d .(ii) behavioural: material-mental:

saFng as activifysayrngas communjcation

He watched them pulling Ehe wa1l down. sensingas activityHe saw Ehem pulling Ehe wall down. sensingasinertprocessing(2) Spatial region: material-relational:The Eroops surrounded Ehe embassy. sPace smolionTrees surrounded Ehe embassy. spaceas relation(3) Symbolization region: verbal-relational:The report indicates they're inEell igent. symbolization/communicationThe report ref lects their inEell iqence. symbolization/identi ty(4) Semiosis region: verbal-mental:He convinced them they'd won. externalsensingHe concluded they'd won. internal s)nng(5) Affection region: mental- relational:He d like them Eo come sensingas nert ProcessHe'd be pleased for Ehem to come. sensingas atfribute(5) Qualitative change: material-relational:His grip weakened changeas eventHis grip became weaker chatgeascomingtobe(7) Representation: material-relational:He is standing for parliament representation as activifyHe stands for honesty rePresentationas (inert)svmbolization

6 To avoidclutter n the igurewe have eftout someclines.For nstance,here s, arguably, clinebetween eceptivepassive)material nd ntensive scriptiveelational nd a similarone betweenreceptivepassive) tfectivemental nd,again,ntensive scriptiveelatbnal.


31/41

-374-

Having exemplified the clines n thesedifferent regions,we will now discuss he firstfwo very briefly.(1 ) There is a behavioural region, with one cline befween material and verbal andanother befween material and mental. The behavioural processes iscussedabove (Section3.2) can be seen n these erms to occuPya semanticsPace long two dines:

(i) a ciine befween material and verbal processeswhich we can gloss as 'saying asactivity'. This would account, among other things, for the abilify of behaviouralprocess to quote:"Oh no, " he crj-ed/sobbed/wept/grumbled/grunEed./murmured etc.(ii) a ciine between material and mental proce*sswhich can be giossed as 'sensingasactivity'. This would explain the presenceof embeddedmacro-phenomena (the act of'them cooking dinner') in both behavioural and mental process of perception(although the behaviourals cannot project meta-phenomena:behavioural - *He iswatching that they are cooking dinner but mental - He sees hat they are cookingdinner. :He i s waEch ing / fee l ing /sn i f f i ng /observ ing [ t f t . * cook ing d inne r ] ] .

(2) There is a region of spatial processes,with a set of verbs that can rerve either inmaterial clauses or in relational one - surround, stretch, r1tn, go, &c. These may beconstrued as doing/ happening - that is, as motion through space.The Medium will moveaiong some vector, which may be specified as a Locative of direction:She ran all the way l9-lhC-Sl-el-I911.They chased her all Ehe way l9-Lhc-q!-aLi9n.

Aiternatively, they are construed as being in space- that is, as location or extent inspace.The Medium does not move along a vector; it is immobile. If there is a Locative ofdirection, it does not specify the destination of a journey, but rather the endpoint ofsomething that extends in space:The road ran al l Ehe way to the stat ion.

While there is overlap in the sets of verbs that can serve in the material and reiationalcontexts - verbs such as run - the fulI 'material' and 'relational' sets are quite distinct.For instance, we might also have she ogged all the utay to the station and the roadextendedall the way to the station, but neither she extanded ll the way to the stationnor the road jogged all the way to the station.One of the passages rom the text discussed in Halliday (197$ illustrates the potentialambivalence within the spatial region:

(1) He rushed to the edge of Ehe water and came back. (2) Onei ther s ide of the open bank Ehe bushes grew th ick ly in Ehef lood; (3) they waded out (4) unt i l a t the i r far thest some ofthe leaves were opening under water; (5) and these bushes leanedover .The spatial processes n (1) are clearly construed as doing, with an animate and self-propelled Actor; and they are thus material. The circumstance to the edgeof the watq isthus a directional Locative. In contrast, (3) and (5) display a tension: the Medium/ Subjectis inanimate and, in principle, immobile; but the verb is one of motion (3) or potentially oneof motion (5) - d. Halliday (op cit., p. 115).


32/41

The clinesare not all of the same Jrye - somearemorediscreet han others,someclearlyinvolve grammaticalmetaphot, and so on ; bu t it is useful to bring them together forprrrpor"r"of discussion.Sometimes ransitivity has beerrdiscussed n terms of a single scaleiu"iri"g from high rransitivitv to low lransitivify (cf.Hopper & Thompson, 1981); ut thisis in fact only one type of consideration.

43 Topology acrossgrammaticalenvironmentsThe exampleswe have given so far have all beenconcernedwith 'grammatical spaces'that are construed fypiogically by system networks within a given grammaticalenvironment, such "r th" cliuse. For-insiance,he spaceof processess constmed in thegrammatical environment of the clause. n the first instance, these environments arecreated,and also separated,by rank - the environment of clauses,grouPs & phrases(and,within these, he primary classes f nominal, verbal,etc.).The basic pnncipie here is anexperiential one: the complexity of experience s construedgrammatically as_.comPosition.Thus goings-on are inteipreted and represented n two stePs - as configurafions ofpro..ri"r, "participants and circumstanc-esclauses)and as their comPonent parts, thepro""rr"r, participants and circumstances hemselves groups & phrases)'Tit distributionbf tn" grammatical system into subsystemsoriginating in different, ranked environmentsmakes"good experiential sense,most of the time. However, it has the effect of dispersinginterpeisonal systemsof affect across hesedifferent grammatical environments. Thus,attituae is manifested n the clauseas adjuncts (unt'ortunately, appily &c') and in thenominai group as pre-modification (attitudinal epithets: unfortunate,poor; happy &c')'There is a sense, hen, in which the sameregion in a topological space (e.g. affect) is'revisited' or, manifested, in different environments in the grammar. In fact, we find thisphenomenon also within the ideationai metafunction as Halliday 1985:306-7has shown -- the 'cyptotypes' of elaboration, extensionand enhancementare manifested withinclauseco.t pter,'es, lauses,prepOsitionalphrases,etc. (see urther below, Section5' Tabled) . In a typological u..oott, there ls no relationship between these differentmanifestations of iheseregions - preciselybecausehe manifestationsare dispersed acrossdifferent environments iri the grammar. All we can do at present is posit a generalizedsystem that we 'abstract' from the specific grammatical environments in which it ismanifested:/ seeTable2.

Table 2: Absracted systemand manifestationsin different environments7 Alternatively, e could bring this systemou t diagrammaticallyy means of a helix, as in theconclusionof Section2 above-

'abstracted'sgstem sgstem'locrtcdngrammaticalnvironment

"lausc "omgl"r. clause: elational preposit ional phrase

elaborating

cxtcnding

cnhencing

elrborriing' i . c . 'cxtcnding'rnd'

cnhrncing'then' &c .

intensive'be'

posscssivehavc'

circumstantial'bc rt ' &c .

elaborat'in9'of'

cxtcnding'with'

enhrncing' a t ' & c .


33/41

- J / U -

We can think of this as a projection of a general, fractal' principle tJrrough he grammar;the particular images projectedwill ai l be similar but there will also be differences naccordancewith the different environments.For instance, n the environment of reiationaltransitivify in the clause,elaboration, extension,and enhancementare interpreted asprocess hat can unfold in time.

5. Extant topologiesAs we might expect, opologicalwork is not unknown in systemicdesaiption, although thas never been given the positive characterisationwe are attempting to formulate here.Hall iday's account of MODALITY and MODULATION as degrees of usuali ty,probability, obligation and inclination presentedabove is one well known example of atopological systemic description complementing a typological one. WithMODAUZATION, POLARITY in propositions s used to bound the semanticspaceunderconsideration,with degree of usuaiity and probabilily defining parametersbetween 'it is'(yes) and 'i t isn't' (no). Similarly, obligation and inclination are used as parametersgrading the semanticspace n proposalsbetween'do' and 'don't'. - SeeSection3 above.Additionai examples inciude Halliday's earlier work (1975:2'12) n modalify andmodulation in which he draws out the parallels between the ways in which the fwosystems divide up a remantic spacewhich at that time he interpreted as deriving fromdistinct interpersonal and ideational metafunctionalsources.

[insertTable V here]Table 3: MODALITY

Also in this general areas of meaning, Halliday 1985: 336 scales congruent andmetaphorical realisations of MODALITY and MODULATION with respect to degree ofsub ectivity / o bjectivity:[insert Table 10(2)here]

Table 4 Degreesof subjectivity/ objectivityHalliday 1985:251 iso provides an ideationai example in scaling phenomena fromThings (least clause-like) to Quotes (least thing-like):

linsert Table 7(15)herelTable 5: From things to quotes

Far and away the most eiaborated example of topological description was mentioned inSection 4.3 above; it is found on Table 9(3) on Halliday 1985:30G307where he projects thelogical parameters of elaboration, extension and enhancement and their sub-fypesthroughout the grammar. The Table is presented as a matrix, with lines dividing thevarious regions into independent cells. In spite of this the cross


34/41

- 3 7 7

5. Environment(con/textual cology)At this point it is important to stopand askwhy topologyhappens? or rather,why isit that fypologically distinct grammatical classes dapt themseives o function in similarways? As with sharksand dolphins, the answerof course ies in the envrronment. In thecaseof language he environment s the socialsemiotic n rvhich anguagePatternsevolve.This means that a topologicaiperspectives an ecologicalone - and that in order to

understandwhy typologiesevolve in sucha way that divergent meanings come to workand look alike we need to look at text in context. For it is contextualised ext that drivesthe grammafical semogenesishat leads to the need for complementary typological andtopologicalorientations n our derriptions.To illustrate briefly the kind of environmentalpressureswe are referring to here,considerthe following text (the first paragraph from My Place , Wheatley & Rawlins, 1'987). t isthe third sentence underlined) which particularly concernsus here.

M y n a m e ' s L a u r a a n d t h i s i s m y p ] a c e . I t u r n e d t e n l a s t w e e k .Our house is the one w i th the f l ag on Ehe w indow . Tonv savs i tshows we ' re on Abor io ina l l and . buE f th ink i t means Lhe co lou ro f t he ea r fh hack homc. Mum and Dad I i ve here Eoo , and Ter r yand Lorraine, and Aunt ie Bev, and Tony and Diane and therr babyD e a n . H e ' s m y n e p h e w a n d h e ' s s o c u t e l W e c om e f r o m B o u r k e ,bu t Dad t .houghE there 'd be more jobs in Ehe c iEy .

This sentence s a dause complex consistingof five ranking clauses. Their logical analysisis presentedbelow:1 d Tony says"p i t shows' /Y we're on Abor ig ina l land,+ 2 a b u t r t h i n k'p iE, means Ehe colour of Ehe earth, back home.

The second ranking clause, it shws, is a verbal process,hypotactically projecting thelocution we're on Aboriginal and; it could have a Receiver(it shws us). The final clauseon the other hand, it means he colour of the arth, backhome, is a relational identifyingprocess- a Token (if ) ^ Process means ^ Value (the colourof the earth, backhomestructure with which the symbolic significance of the Aboriginal flag is construed. Notehowever that both clausesare concernedwith the meaning of the flag: it shows exploresits significance verbally, while it means explores it relationally:'signifier' 'signified''i t- I i-hc f 1arr ) shows we ' re on Aboriginalv - - - - - s y ,

l a n di t ( the f lagr) means Ehe colour of Eheear th - back home

This is in fact the 'symbolization as communication/symbolization as identity'topological parameter noted in Section 4.2 above. And the text illustrates the way inwhich the verbal and reiational poles of this parameter can be marshalled textually to dodosely related work. Indeed, this is just the type of example that makes typologicalanalysis difficult - context so powerfully construes the different structures as the samethat it hard (especially for apprentices) to see that grammatically how they are distinct

COMMUMCATIONIDENTITY


35/41

- B -

There are pientv of other eramples of this kind in the literature which demonsffate thekinds of con/te;


36/41

- 3 7 9 -

c l c s e a n o . , v h v ? T h e s e a r e s c m e o f 1 : h e f u e s - - i o n s - - c c e a n s w e r e o- . F a r < ^ a a n r o rHere again we seea particular discoursepushing different areasof the grarrunar to dosimilar work, rvith nominalisation eeding the compiementaryExistent and V{edium roles.In related research nto the discourseof geography, the same researchersexplored the

processesby which technicai terms are createdand defining. Wignell, lv{artin & Eggtns19871,990note that elaborationat word and group ranks and identif ication at clauserankare ail integral to the processof definition (cf. Halliday's Table 9(3) reproduced as Table 5above).word complex:

goup compiex:

relational identifying clause:

n r ^ f n 7 ^ :i

the f ina l leve laI

1 q i n a l o - c o l I o r l a n i m a l q \a

/ F r n n h i r ' l o r r a l ? \\ v - v r . . + v J I^t hey a re ca l l - ed conver te rsToken Process Value

The co-operation between these different areas of the grarrunar is illustrated in thefollowing text ('elaborations'are underlined). . . A E E h e n e x t l e v e l ( t r o p h i c l e v e l 2 ) a n i m a l - s e a t E n e s e o l a n t s( h e r b i v o r e s ) o r e a c o t h e r a n i m a l s ( c a r n i w o r e s ) ; E h e s e a n i m a i sa r e t h e c o n e u m e r g o f t h e e c o s v s t e m . A E t h e f i n a l l e w e l -1 l - r n n h i a l o r r o l ? ' l h a n ' l - o r i e f r r n a i , a n r l r - o r f , a n n r a ) l - . 7 - ^ / s . i n r r l c -4 e s 5 r r r yc e l 1 e d a n i m a l s ) b r e a k d o w n d e a d a n j - m a l s a n d p l a n c s a s w e l l a sEhe wasEe produc ts o f the consumers . These Ehe lz conver t in tou s e f u l e l e m e n E s w h i c h a r e t a k e n u p o n c e m o r e b 1 z t h e r o o t s o fp l a n t s a n d s o t h e v a r e c a 1 1 e d e o n v e r t e r g .

To date, there appears to be no systemic documentation, either phylogenetically orontogeneticaily,of the ways in which contextualpressures f this kind brings pressures obear on the grammar which force t b elaborate ts options n ways that divergent choicescome to resembleeach other. But this is obviously a rich and exciting frontier for systemicdescription which will bear critically on our current rather limited understanding ofsemogenesis.

T.ImplicationsWe began by surveying the typological resourcesof systemic theory; in particular/ wetried to show how they are 'multidimensional' becauseof stratification, simultaneousmetafunctions, and so on. We then suggested that there are still paradigmatic relationsthat are hard to capture and turned to a topological perspective as a complement" Inopening up a new descriptive frontier of this kind, it is hard to avoid a higher proportionof queries than insights. Nevertheless, here are a number of implications associatedwithproposing complementary typological and topological orientations which we will brieflyreview here.(i) One thing we have noted in our teaching of the grammar is than grammar learners areborn topologists. A very high proportion of the 'mistakes' in student analysesof text haveto do with theh tendency to background the criteria motivating particular classificationsand foregrounding alternative criteria which would tend to ground a complementarytopological perspective. This is even more true where texts push different areas of theglarrunar in similar ways such as those illustrated in Section6 above. This makes student


37/41

- 3 8 0 -

assignments fertile ground for topoiogicalenquirv, the resultsof which can be used toaileviatesome of the frustration they feel rvhenbeing down-gradedby narrow mindedtypologrstssuchas ourselves.(ii) Another way in which topologicaiwork should prove useful hasto do with resolvingtensionsabout which criteria to use as point of departure n system networks. Arypracticing systemicist viil be familiar with the problems when writing a system networkof knowing where to start, which problem always resolves tself when made explicit intothe issue of which criteria to deploy at primary, which at secondary delicary and so on.Given qrpology and topolog'y as complementary descriptive strategies, resources areprovided which make it possible found networks on certain criteria and use others toestablish parameters within some relevant semantic space that might otherwise beobsorred.(iii) Topological considerationsare also highly relevant to argumentation and debate.As more than one systemicisthas complained (e.g. Berry,1989), argumentation amongsysternicistsoften suffers from the lack of explicit attention paid to criteria for motivatingfypological descriptions; distributing criteria across typologicai and topological

perspectives should encourage systemicists o make their criteria clear. Beyond this,systemic descriptions are often criticised uninsightfully on topological grounds. Theargument runs along the lines of "I can seea dimension of agnation other than yours (whichI actually can't see); hereforeyou are wrong.'' Huddleston's1988 eview of Halliday 1985is full of argumentation of this kind - e.g. his critique of Halliday's analysis of clausessuch as after they had debatedt'or an hour in aftn they had debatedt'or an hour theyshook ands as hypotactically dependentclauses n a clausecomplex instead of as dausesembedded in other clause.Huddleston wants to capture the agnation between after theyhad debated hey shookhands and after the debate hey shookhands; bu t he saysnothing about the agnation (captured in Halliday's analysis) between after they haddebated.'or an hour they shookhands and they debatedt'or an hour, then they shookh a n d s .(iv) We have considered topological parameters within the framework of a singlelanguage, English. But topological considerationsmay prove even more useful in work onlanguage typology, which has long been troubled by the problem of looking for categories(i.e. typologically defined classes) in languages (and usuaily finding them; see Martin1983 or a discussion of this problem). A topological perspectiveswould allow us to designparameters for comparison within various kinds of semantic space. And topologicalpararneters n one language might resolve into typologicai patterns in another. Recall forexample the discussion of the symbolization parameter connecting verbai and relationalprocessesn Section3 above. br Tagalog the normal way b quote is not in fact to project as

in English, but rather to establish a Token-Value relation between a locution and anominalised verbal process. Our topologrcal perspectiveon English TRANSffIVITY tellsus in effect not to be sqprised.ang sagoc n lya .TM aruwer s/heValueher answer

'TaposfinishedToken

n :fruition"Ifs over now""It's over now", she said.

(v) The question of ontogeny and phylogenesis has already been raised above. Let ussimply mention here an example, noted by Halliday 1985:128 n his discussion ofbehavioural processes.He notes there that behavioural processes n English used 'present'rather than 'present in present' ense b constmct ongoing action and that becauseof this westill find some alternations such as Why do you laugh? alongside Why are you laughing?One may say in this context that behaviourals are drifting in the direction of material


38/41

- -

processeslong he opologicai arameter e estabiishedn Sectron . Indeed, t mav oniyte in modern nglisir itrat he tension veconsideredvouidarise. The rvayTENSI used towork we might iell haveclassified ehaviourals ith mentalsand verbals probabiynotrepresentit i ttr"* as a distinct classat ail). Putting this boldlY, we could claim that,"*og".,"siJ proceeds long topologicalparameters,nd that certainkinds of changemaymean-thatweneed to reconsider ur typologrcal escriptionrom time to time.(yi) Finally we should mention Lemke'sown suggestion oncerning he relevanceoftopologicalperspectiveso language earning. Lemkepoints out that taking topologicalreiationsbetweengenresnto aciountailowsus to Plan eachingprograrnsn which we aredear about the rvJys n which teachingonegenremight providea gogq basis or movingonto another topiogically relatedone. Similarly n grammar teaching,a toPoiogicalperspective hould make-it easier o make explicit the connections etweensystemicgrammatical descriptions and the kind of learner oriented gralrunars envisioned by

iropor,".,ts of the functionalnotional syllabus e.g. Wilkins 1976). At presentsystemicgtu*t t"rr differentiatenotionsand functions hat may need to be brought together forESL/EFTpurPoses.Our hope is that by pursuing work along a number of these ines of enquiry ra/ecalformulatb our charait.ri*tiotr of topologicalanalysis n more positive terms. This willmeandevisingsuitable orms of repr-esentation,rganised roundmetaphorsof spaceandworking out proceduresor motivatingparameters ndbounding relevantfypesof semioticspace."In the short term we needto-work hard on improving our understandingof thectmplementary relationshipbetween ypologicaland topologicalorientations so that assystemic inguistswe have a stto.,gergrp on how making typology the centerpieceof ourtireory and iescriptions selects r f-oregroundsertainparametersn a systemand leavesothersmore implicit.In conciusion, et us just note that just aswe haveexaminedparadigmaticorganization intermsof topology,*"."r, alsoexploresyntagmatic rganizationalong similar lines. Forinstance, trb exfriential constituentstmctureof the dausesays here s a processelement,one or morepuiticipar,ts,and possiblyaisocircumstances.rom a topologicalpoint of view,the process nd thi medium ire moit closelyassociatedHatliday, 1985) they form thecore or nucleus- whereasother participantsand circumstances re lesscloselyassociated.


39/41

- 3 8 2 -

ReferencesBateman, . A. 1989.Dynamic svstemic-funcfionai rammar:a new frontier. Word 40.1-2(Systems,Structures and Discourse:seiectedpapers from the Fifteenth lnternationalSystemicCongress).253-286.Berry, M. 1989. They're all out of step exceptour ]ohnny: a discussionof motivation (or thetait< of i t) in systemic l ingusitics. OccasionalPapers in Systemic Linguistics 3.Departrnent of English Shrdies,University of Nottingham. 5-58.Eggins, S, P Wignell & J R Martin. 1987. The discourse of historv: distancing theieco"erable past. Working Papers in Linguistics 5: Writing Project - Report 7987.Department of Lingusitics, University of Sydney. 1987.5G176.Fawcett, R. P. 1980.Cognitive linguistics and social interaction: towards an integratedmodel of a systemic functional grammar and the other comPonentsof a communicatingmind. Heidelberg:JuliusCroos.Kress, G IEd.l 1975. Halliday - System and Function in Language. London: OxfordUniveristy Press.Halliday, M. A. K. 1961. Catgoriesof the theory of grammar. Word 17.3.241'-292.Halliday, M. A. K. 1958. Notes on transifvity and theme in English: Part 3' Journai ofLinguistics 4.2. 179-?75.Hailiday, M. A. K. 1970. Functional diversity in language,as seen rom a consideration ofmodaiity and mood in English. Foundations of Language 6.3.322-361. [Reprinted in C

Kress tEd.l Halliday - System and Function in Language. London: Oxford UniveristyPress1976lHalliday, M. A. K. 7978. Language as a social semiotic: the social interpetation oflanguage and meaning. London: Edward Arnold.Halliday, M. A. K. 1,982. The de-automatization of grammar: from Priestiey's 'AnInspector Calls'. J M Anderson [Ed"] LanguageForm and Linguistic Variation: papersdedicated to Angus Macintosh. Amsterdam:Benjamins.129-159.Halliday, M. A. K. 1984.Language as code and language as behaviour: a systemic-

functional interpretation of the nature and ontogenesisof dialogue. R Fawcett, M A KHalliday, S M lamb & A Makkai [Eds.] The Semioticsof Language and Culrure: Vol 1:Languageas SocialSemiotic.London: Pinter.3-35.Hailiday, M. A. K. 1985. An Introduction to FunctionalGrammar. London: Edward Arnold.Hopper, P. & S.Thompson. 1980.Transitivity in grammar and discourse. Language,55.Huddlestotr, R. 1988. Constituency, muiti-functionality and grammaticalization inHalliday's Functional Grammar. Journal of Linguistics,24: 136-74.

Lamb, S" 1971"The crooked path of progress n cognitive linguistics" Monograph Seri191Languages and Linguistici 24. 9972{" lreprinted in A Makkai & D G Lockwood [Eds.]Realin[s in Sfratificational Linguistics. Universify, Alabama: The Universify ofAlabama Press. 1973.12-331


40/41

Lemke,J. L. no date. The topologyof genre: ex t structures nd text frpes. UnpublishedPaPer.

Lemke, . L. 19g2.Semiotics nd Education.Toronto:VictoriaUniversity (TorontoSemioticsCircie Monographs,working Papers& Prepublications).Lockwood, D. G. 7972. Introduction to stratificationalLinguistics. New York: Harcourt,

Brace, ovanovich.Martin, J. R. 1983. Participant identification in English, Tagalog and Kate. AustralianJournalof Linguistics3.1. 45'74.Martin, J. R. 1985. Process nd text: rwo aspects f semiosis.J D Benson& W S Greaves[Eds-]Systemicperspectives n Discourse ol. 1: seiectedheoreticalpapers rom the 9thinternatinal Systemii Workshop.Norwood, NI: Ablex. 248'274.Martin, J. R. 1986.Grammaticalisingecology: he politics of baby sea-ls nd kangaroos'T

Threadgold, E A Grosz, G Kressk M A K HailidaY: !^tgt1ge, Semiotics,deology'

Sydneyisydney Association or Studies n Societyand Culrure (SydneyStudies n Societyand Cultnreil.225-26.Matthiessen,C. 1988. Representationalssuesn systemic unctional grammar. J D Benson& W SGreavesEds.]Systemic unctionalApproacheso Discourse.Norwood, N J:Ablex'73G175.Matthiessen,C. 1989. Lexicogtammatical artograPhy:English systems.Department ofLinguistics, University of Sydney:mimeo'Matthiessen,C. & M.A.K. Halliday. 1989.SystemicFunctional Grammar. In J. Ney & F'peng (eds),Current approacheso syntax.London & Amsterdam:Whurr & Benjamins'Forthcoming.poynton, C. lgg4. Names as vocatives: orms and functions. Nottingham LinguisticsCircular 13 (Special ssue on SystemicUnguistics)' 1-il'poynton, C. 1985. Languageand Gender:making the difference-Geelong,Vic.: DeakinUniversity Press nepu6[sfied London:Oxford University Press.1989]Trew, T. lg1g. 'What the paperssay': inguistic variation and ideoiogical difference. RFowler, B Hodge, G Kress& T Trew. tanguageand Control. London: Routledge& KeganPaul. 717-756.Ventola, E.7gg7. The structure of socialInteraction:a systemicaPProach o the semioticsof serviceencounters.London: Pinter.wignell, P., J. R. Martin & s. Eggins._19!7.The discoulseof geograply_:ordering. andeiplaining the experiential *oitl. Working PaPers n Linguisti!: 5 - Writing -Pt-ojq1-neport tg\2. Department of Linguiltics, U*versity of Sydney. &116. lRepublished inLinguisticsand Education 1.41990.359-3921Wilkins, D.7975. Notional Syllabuses.London:Oxford University Press.


41/41

Lemke, . L. no date. The topoiogyof genre: ex t slrucfures nd text fvpes. UnpublishedPaPer.

Lemke,J. L. 19g2.Semiotics nd Education.Toronto:VictoriaUniversity (Toronto SemioticsCircie Monographs,working Papers& Prepublications).Lockwood, D. G. 1972. Introduction to shatificationalLinguistics. New York: Harcourt,

Brace,Jovanovich.Martin, J. R. 1983. Participant identification in English, Tagalog and Kate. AustralianJournaiof Linguistics3.1. 45'74.Martin, J. R. 1985. Process nd text two aspcts f serniosis.J D Benson& W S Greaves[Eds.]SystemicPerspectives n Discourseol. 1: seiectedheoreticalpapers from the 9thinternatinal Systemii Workshop.Norwood, N: Ablex. 248'274.Martin, J. R. 1985.Grammaticalising cology: he politics of baby sealsand kangaroos'T

Threadgold, E A Grosz, G Kress-& M A K HallidaY: !^"g"1ge, Semiotics,deology'

Sydneyisydney Association or Studies n Societyand Culture (SydneySfudies n Societyand Culture$.2?,5'258.Matthiessen,C. 1988. Representationalssuesn systemic uncfional grammar. i D Benson& W SGreavesEds.]Systemic unctionalApproacheso Discourse-Norwood, N J:Ablex'73G775.Matthiessen,C. 1989. Lexicogrammatical artograPhy:English systems.Deparfment ofLinguistics, University of Sydney:mimeo'Matthiessetr,C. & M.A.K. Halliday. 1989.SystemicFunctional Grammar. in J-Ney & F'Peng (edsi,Current approacheso syntax.London & Amsterdam:Whurr & Benjamins'Forthcoming.poynton, C. 1984. Names as vocatives: orms and functions. Nottingham Linguisticsiircular 13 (Special ssueon SystemicUnguistics)' i-&'poynton, C. 1985. Languageand Gender:making the difference.Geeiong,Vic.: DeakinUniversity Press nepuLtsfied London:Oxford University Pre*s.1989]Trew, T. lg7g. 'What the paperssay': inguistic variation and ideological differ:1c-e. RFowler, B Hodge, G Kress'&T Trew. tanguage and Control. London: Routledge& KeganPaul. 117-756.ventola, E.7gg7. The structure of social Interaction:a systemicaPProach o the semioticsof senviceencounters.London: Pinter.wigneil, P., J. R. Martin & s. Eggins.,19p7.The discourseof geograp\y; ordering. andeiplaining the experiential *oiti. Working PaPers n Linguistig: 5 - Writing,Pt-ojqf -neport tg\2. Department of Linguistics, Uriversity of Sydney. &'776. [Republished inLinguisticsand Education 1.41990.359-3921

Documents

BC-1992-Systemic Typology and Topology