Verb classes and valency alternations (NSM approach), with special reference to English physical...

Preview:

Citation preview

[Revised MS] In A. Malchukov & B. Comrie (eds.). In press/2015. Valency Classes in the World’s Languages. Berlin: Mouton de Gruyter. Verb classes and valency alternations (NSM approach), with special reference to English physical activity verbs Cliff Goddard This study examines five English physical activity verbs from the Leipzig Valency Patterns Project list (eat, pour, dig, carry, cut). It proposes detailed semantic explications for the basic activity-in-progress meanings of these verbs and shows how these can be transposed into perfective uses. It examines and explicates 11 alternations (specialized constructions) involving these verbs, showing in each case exactly how these constructions are related to the base semantics of the verb. The general picture is that the specialized constructions are quasi-derivational in nature: the primary or semantically basic sense of the verb is embedded in a more elaborate configuration containing additional semantic material. Often much of this additional material is modeled on the semantics of verbs that belong to different semantic types (lexicosyntactic blending), but it can be partly idiosyncratic or non-predictable. Each specialized construction represents a kind of “word in construction” polysemy.

1. Introduction 1.1 Lexical semantics and alternation phenomena

Many linguists see the verbal lexicon as organized into numerous classes and subclasses, each sharing certain more or less distinctive semantic and syntactic properties. To a large extent this outcome can be traced to work begun by Beth Levin (1993) in her book English Verb Classes and Alternations. This is somewhat ironic, because Levin was insistent that “the important theoretical construct is the notion of meaning component, not the notion of verb class” (Levin 1993: 18). The verb class concept, she said, was an “artificial construct” useful only for formulating generalizations at an approximate level. A decade later, with scores of further studies in hand, Levin and Rappaport Hovav (2005) reiterated that alternations cannot be explained in terms of discrete verb classes. There are too many cross-cutting and overlapping alternation patterns for that.

Such data lead to the conclusion that is the elements of meaning that define verb classes that are most important, and that verb classes themselves are epiphenomenal … even if they might be useful in the statement of certain generalisations. Therefore, advances in the study of argument realization regularities require isolating those semantic components which ultimate determine them. (Levin and Rappaport Hovav 2005: 16)

Isolating the relevant semantic components of course requires a “well-motivated theory of lexical semantic representation” (Levin & Rappaport Hovav 2005: 23). A sound methodology for lexical semantic analysis is crucial for another reason as well, namely, the connection between alternations and lexical polysemy. As will shown in the present chapter, valency alternations are properties that belong to lexical units, rather than to whole lexemes. A clear picture of valency phenomena, both within and across languages, therefore requires careful attention to lexical polysemy. Further complications follow from the intimate association between valency and alternation phenomena, on the one hand, and aspect and event composition, on the other. The whole area demonstrates the integration of lexical semantics and grammatical semantics. As Wierzbicka (1988: 8) put it, in a work that anticipated current themes in construction grammar and cognitive linguistics:

Cliff Goddard

2

There is no such thing as ‘grammatical meaning’ or ‘lexical meaning’. There are only grammatical and lexical MEANS of conveying meaning — and even here no sharp line can be drawn between them. (Wierzbicka 1988: 8)

Despite the implications of the term ‘alternation’, the phenomenon is not simply a matter of the arguments of a verb being “realized” differently or assuming different argument roles. As Dowty (2000: 110) put it, the “main linguistic phenomenon that ought to be of interest” is that the alternate forms “serve to convey significantly different meanings”. What we are dealing with under the rubric of valency alternations are lexicosyntactic constructions for expressing specialized meanings – meanings that cannot be captured without semantic analysis at a very fined-grained level of resolution1. 1.2 Semantic approach

The current study is conducted within the Natural Semantic Metalanguage approach (Wierzbicka 1996; Goddard & Wierzbicka 2002; Goddard 2008, 2011, 2012; Peeters 2002; Ye 2010, in press; Gladkova 2010; Bromhead 2009, 2011; Wong 2005, 2014; Levisen 2012; and other works). Though still regarded by many as controversial, it is arguably the most highly developed system of lexical semantic analysis on the contemporary scene. For present purposes, it is not possible or necessary to justify its every tenet and claim, but only to delineate its assumptions and operating procedures. The NSM method of meaning description is reductive paraphrase. Its primary tool is a constrained vocabulary of 65 posited universal semantic primes, which have been arrived at after a lengthy program of semantic research that began with Wierzbicka (1972). Semantic primes are defined as the terminal elements of reductive paraphrase analysis, i.e. meanings that cannot be paraphrased in simpler terms. The current inventory of primes is displayed in Appendix 1, in their English versions. Examples include: SOMEONE, SOMETHING~THING, PEOPLE, DO, HAPPEN, WANT, THINK, KNOW, GOOD, BAD, BECAUSE, CAN, and LIKE~AS. Evidence suggests that semantic primes are likely to be present as the meanings of lexical units, i.e. as discrete meanings of words or word-like elements (bound morphemes or phrasemes), in all or most languages. Semantic primes have certain inherent combinatorial (syntactic) properties that follow from their individual meanings. The syntax of the metalanguage is significantly complex, including clausal operators, such as CAN and NOT, and allowing temporal and clausal adjuncts, such as AT THIS TIME, AFTER THIS, and BECAUSE OF THIS. It also recognizes that some primes have valency options and complementation possibilities. The most complex valency properties are found with the predicate primes. There are 17 of these primes, making it impractical to review them here in any detail, but the valency frames available to HAPPEN and DO are displayed in Table 1. Because they are given in their English versions they of course include certain English-specific features, e.g. the prepositions to, with, and about, but the same frames can be realized in other languages using different linguistic devices, such as case-marking, postpositions, verb serialisation, or purely by word-order.

1 The terminology of “alternations” and “argument realisation” originated in generative linguistics. Though not necessarily inimical to clear discussion, in my view it is not particularly helpful either. The situation is analogous to that surrounding the term “movement”, which is still used as a handy descriptive label by many linguists who do not accept a transformational account of ordering phenomena. For the purpose of the Leipzig Valency Patterns Project, the key term is “valency alternation” as defined in the project manual.

Verb classes and valency alternations (NSM approach)

3

Table 1. Valency frames for semantic primes HAPPEN and DO

something HAPPENS [minimal frame] something HAPPENS to someone/something [undergoer frame] something HAPPENS somewhere [locus frame] someone DOES something [minimal frame] someone DOES something to someone else [patient1 frame] someone DOES something to something [patient2 frame] someone DOES something with something [instrument frame] someone DOES something with part of the body [body-part frame] someone DOES something with someone [comitative frame]

Every prime has its specified individual mini-grammar of combinatorial possibilities. It should be noted that exponents of some primes have a narrower range of options in NSM than they have in “full” English. For example, in ordinary English one can speak of ‘doing something about something’, but this (highly English-specific) option is not available in NSM. Together, the semantic primes and their grammar of combination constitute a mini-language which is an ideal tool for semantic analysis within a single language and across languages: hence the term “natural semantic metalanguage”. Versions of this metalanguage have been documented for French, Spanish, Russian, Polish, Danish, Finnish, Chinese, Korean, Japanese, Malay, Lao, Mbula (PNG), Amharic, Arabic, East Cree, and a variety of other languages (see individual chapters in the edited volumes: Goddard & Wierzbicka 2002; Peeters 2006; Goddard 2008; also Gladkova 2010; Habib 2011; Levisen 2012; Vanhatalo, Tissari & Idstöm 2014). When describing complex meanings in any language, the goal of NSM lexical semantic analysis is to arrive at a well-evidenced reductive paraphrase (an ‘explication’) for each discrete sense of the word (or other expression) under consideration. The explication may be framed exclusively in semantic primes or it may include some ‘semantic molecules’, in addition to semantic primes. The term ‘semantic molecule’ refers to certain non-primitive word-meanings which function as units in the structure of many other concepts from diverse domains across the lexicon (Wierzbicka 2009a; Goddard 2010). Examples of productive molecules of English include: ‘hands [m]’, ‘sharp [m]’, ‘water [m]’, ‘ground [m]’, and ‘hold [m]’. The notation [m] identifies a word as a semantic molecule. Because they are ultimately decomposable into primes (without circularity), there is no violation of the non-circularity principle. Some explications make use of complex terms of a different kind: derivational bases. For example, the explication for a derived word like illness includes the base element ‘ill [d]’, where the notation [d] indicates a derivational base (Goddard & Wierzbicka 2014). Unlike semantic molecules, which range widely across many domains of the lexicon, derivational bases have a localized and specialized provenance. In §3, we develop a quasi-derivational approach to aspectual transposition and to alternations (specialized constructions). 2. Approaching valency and alternation phenomena, NSM style In NSM semantics, explications for words which are semantically and grammatically similar are expected to follow a consistent organizational format or pattern. Such organizational formats are termed semantic templates (Goddard 2012). For verbs, the top-level components of the template constitute the so-called Lexicosyntactic Frame. These components account for much of the morphosyntactic behavior of a given verb, as they identify the core participants and their semantic roles, inherent aspect, and, often, components of volitionality and control. In §2.1 we review a range of Lexicosyntactic Frames for various verb classes. In §2.2 we outline the remaining parts of the template structure for physical activity verbs in particular.

Cliff Goddard

4

2.1 Lexicosyntactic Frames

Grouping verbs according to shared Lexicosyntactic Frames is one analogue in the NSM framework for the notion of “verb class”. Consider the configuration shown in [a]. The first line (‘someone does something to something for some time’) closely matches the idea of a transitive activity, while the second line depicts a concurrent “ongoing effect” on an object. If the agent exercises significant control over the ongoing effect (as, say, with cut), this can be captured by the additional component ‘as this someone wants’. An instrument can be provided for by adding the line: ‘this someone does it with something’. Variants of the configuration in [a] constitute the Lexicosyntactic Frames for many transitive physical activity verbs, such as eat, drink, cut, peel, and grind. a. someone does something to something for some time because of this, something happens to this something during this time (as this someone wants) (this someone does it with something) Many verbs of locomotion involve an actor undertaking a localized activity (doing something somewhere) with the concurrent result being controlled movement in the place in question, as shown in [b]. This is the Lexicosyntactic Frame for verbs such as walk, run, crawl and climb (Goddard, Wierzbicka & Wong in press). b. someone does something somewhere for some time because of this, this someone is moving in this place during this time as this someone wants The examples so far have involved DO in combination with the open-ended durational expression ‘for some time’. It is equally possible for DO to appear with the time adjunct ‘at this time’, to which can be added ‘in one moment’ to depict punctuality. Many action verbs involve top-level components like these, followed by one or more components that spell out a subsequent result. This can be an effect on a person, as in [c], for verbs like hit and kill, or an effect on a thing, as in [d], for a verb like break (Goddard & Wierzbicka in press). c. someone does something to someone else at this time because of this, something happens to this someone else at this time (it happens in one moment) d. someone does something to something at this time because of this, something happens to this something at this time (it happens in one moment) The result can also be a change in the location of an affected theme, as in [e] or variants thereof, for verbs like put and throw, or a period of ongoing displacement, as in [f], for carry. e. someone does something to something at this time because of this, after this, this something is not in the place where it was before f. someone does something to something for some time because of this, this something is not in one place during this time, it is in many places at many times The top-most predicate in the examples mentioned so far has been DO; that is, they have all concerned “verbs of doing”. Needless to say, the other primitive predicates such as HAPPEN and SAY (among others) can also figure in comparable roles. For example, many verbs of “deterioration” begin with the components in [g] (Barrios & Goddard in press). Many speech-act verbs, such as ask, tell, and promise, begin with the components in [h] (Goddard & Wierzbicka 2014: Ch 6). g. something bad happens to something for some time because of this, after this, this something is not like it was before h. someone says something to someone else at this time this someone wants something to happen because of it

Verb classes and valency alternations (NSM approach)

5

Many other combinations and configurations are possible (Goddard & Wierzbicka in press). 2.2 Semantic templates for physical activity verbs

As mentioned, verbs of different kinds can have different template structures. Previous NSM work on physical activity verbs (Goddard & Wierzbicka 2009; Ye 2010: cf. Goddard 2012) has established that these verbs have a four-part template:

LEXICOSYNTACTIC FRAME PROTOTYPICAL SCENARIO (incl. prototypical actor’s motivation) MANNER (incl. incremental effect on the object) POTENTIAL OUTCOME

Little more needs to be said about the Lexicosyntactic Frame for activity verbs. The only thing to be added is a reminder that as an explicatory strategy, NSM analysts begin by identifying a semantically basic frame for verbs of any given type. For physical activity verbs, this basic frame is the progressive: the activity-in-progress. In other words, we begin the process of explication with sentence frames such as: Someone is eating (carrying, cutting, etc.) something. The Prototypical Scenario has the following form. The dashed lines represent a description that typically includes the motivation/intention of the prototypical actor, as well as, in some cases, relevant physical qualities of the object. – at many times when someone does this (to something), it is like this: – – – – – – – – Unlike research in extensionalist approaches (e.g. Majid & Bowerman 2007), NSM researchers contend that people conceptualize human activities not solely in terms of their physical characteristics, but partly by reference to the prototypical actor’s motivations. The Prototypical Scenario proposed for English walk, for example, states that a person often does something like this when they want to be somewhere after some time, not far from the place where they are (Goddard, Wierzbicka & Wong in press). Obviously, this does not imply that people only ever walk with this motivation: one can walk for exercise or for pleasure, or for other reasons. Complex physical activity verbs (pour, cut, grind, etc.) have a richer cognitive structure than locomotion and other routine activities, because they prototypically involve something like conscious intention: an actor forming a “preparatory thought” directed towards changing the current state of some object. The Prototypical Scenario for cut, for example, involves wanting something not to be one thing anymore but instead to be two things, and, as well, wanting to control the separation process with some precision (Goddard & Wierzbicka 2009). The next section of the template is Manner. It is introduced by a component that can take either of the two forms below. The second, more complex, form is appropriate for an iterative structure of repeated actions, e.g. for verbs like chop, grind, peel. – when someone does this (to something/somewhere), it happens like this: – – – – – – – – – when someone does this to something, the same thing happens many times it happens like this: – – – – – – – – For bodily physical activities like run, eat and drink, the Manner section describes how the parts of the body are used. For activities like cut, chop, and grind, there is a description of an instrument and how it is used, which can involve several interrelated sub-events. Capping off the Manner section is a component that describes the localized or “incremental” effect as the activity is carried out. For example, the incremental effect of running is that the body moves progressively forwards; the incremental effect of cutting is that the surface of the object is changed by contact with the moving sharp edge of the instrument.

Cliff Goddard

6

The final section of the activity-in-progress template is a Potential Outcome, i.e. an indication of what the final result can be if the activity continues long enough. This component is introduced as follows: – if this someone does this for some time, because of this, after this, ... From a logical (or teleological) point of view, the Potential Outcome component completes the explication by linking the projected process with the motivation of a prototypical actor. This is no accident. As Goddard and Wierzbicka (2009: 88) put it:

given the purposeful, goal-oriented nature of human action, it is only natural that characteristic activities which are prototypically motivated towards certain ends are potentially effective vis-a-vis those ends. In other words, it is natural that there should be a correspondence between the desideratum in the prototypical motivational scenario and final section of the schema: POTENTIAL OUTCOME (Goddard & Wierzbicka 2009: 88)

As we will see in a moment, the Potential Outcome of an imperfective provides the default for an actual outcome when the verb is transposed into a perfective form. Explication [A] below illustrates how the various sections fit together, with a full explication for eat in its activity-in-progress sense. Obviously one could spend a good deal of time discussing and justifying the fine details but this is not necessary for present purposes, and in any case has been done elsewhere (Wierzbicka 2009b; Ye 2010; cf. Goddard 2011: 278-285).

Explication [A]: Someone is eating1 something. someone is doing something to something for some time because of this, something is happening to this something at the same time

LEXICOSYNTACTIC FRAME

at many times when someone does this to something, it is like this: – this something is not something like water [m] – this someone wants this something to be inside their body

PROTOTYPICAL SCENARIO

when someone does this to something, the same thing happens many times MANNER it happens like this: – this someone does something to this something with the hands [m] – because of this, after this, some of this something is for a short time inside this someone’s mouth [m] – after this, this someone does something to it with some parts of the mouth [m] – because of this, something happens to it at this time – after this, this someone does something else to it with the mouth [m] – because of this, after this, it is not inside this someone’s mouth [m] anymore, it is somewhere else

inside this someone’s body for some time

MANNER + EFFECT (HOW IT HAPPENS)

if this someone does this for some time, after this, all of this something can be somewhere inside this someone’s body for some time

POTENTIAL OUTCOME

The level of detail of this, and subsequent, explications (particularly in the Manner +Effect section) may surprise some, and it may be asked: How is the appropriate amount of detail determined? There is no mechanical decision procedure to decide this question. Semantic analysis (see Goddard 2011 for a textbook treatment) is a painstaking process that involves protracted experimentation with successive versions of an explication, having regard to the range of distribution of the word, its implications and entailments, and its relationships with other similar words in the language in question. Valuable semantic clues can also be garnered from comparisons with similar words in other languages. In relation to the Manner + Effect section, it appears to be a semantic characteristic of physical activity verbs that they involve detailed manner specifications. Many other types of verbs, such as verbs for physical acts and for speech-acts, have little or nothing in the way of manner specification (cf. Goddard & Wierzbicka 2009, in press). 2.3 Aspectual transposition: from activity-in-progress to perfective

How are perfective (and other tense/aspect) forms for such verbs explicated? This question has relevance for present purposes because many alternations (specialized constructions) for physical activity verbs appear to be found preferentially in perfective contexts. Briefly, the perfective is produced by a process in which the activity-in-progress template is mapped across to a related perfective template by way of three correlated transpositions. First, the

Verb classes and valency alternations (NSM approach)

7

initial component of the Lexicosyntactic Frame is presented as a temporally localized act (‘someone X did something to something at this time’), without any reference to its duration. Second, the manner in which the event occurred is described in terms of the activity-in-progress meaning. To put it another way, even a perfective use of an activity verb incorporates a reference to a “process”, and this process is depicted by means of the activity-in-progress meaning, i.e. the already explicated, semantically basic sense of the verb. For example, to say that someone ate something implies that this person was eating it for some time. Third, the perfective construction indicates a result: the achievement of a particular outcome. By default the outcome is based on the Potential Outcome in the activity-in-progress sense. The primary activity-in-progress meaning appears marked with [d] to identify it as a non-primitive term with a role akin to that of a derivational base.

Explication [B]: Someone ate1 a hamburger. someone did something to something (a hamburger) at this time LEXICOSYNTACTIC FRAME it happened like this: this someone was eating [d] it for some time MANNER because of this, after this, all of this something was somewhere inside this someone’s body OUTCOME

This general structure corresponds rather closely to standard descriptions of perfective as representation of an event viewed in its totality, without attention to its “internal temporal constituency” (Comrie 1993), or as a “completed event”. Perfectives of physical activity verbs generally allow an additional resultative expression to expand or further specify the achieved result. For example, perfective cut implies that the thing which has been cut is ‘not one thing anymore’, but this can be augmented with more specific information, e.g. She cut the apple into four pieces. Likewise, perfective carried implies that the thing which has been carried is no longer in the place where it was before, but more information can be added to specify where it wound up, e.g. He carried the boxes to the car. The explications below show how this works semantically.

Explication [C]: She cut1 the apple into four pieces. she did something to the apple at this time LEXICOSYNTACTIC FRAME it happened like this: she was cutting [d] it for some time MANNER because of this, after this, this something was not one thing anymore it was four things

OUTCOME

Explication [D]: He carried1 the box to the car.

he did something to the box at this time LEXICOSYNTACTIC FRAME

it happened like this: he was carrying [d] it for some time MANNER

because of this, after this, this something was not in the place where it was before it was in the place where the car was

OUTCOME

Interestingly, the added resultative information can have implications for the event structure being described. For example, cutting does not have to be iterative (one can cut something in a single movement), but to cut something into four pieces requires at least two cutting movements. 3. Valency alternations of physical activity verbs Various subclasses of physical activity verbs can be identified for English on the basis of their Lexicosyntactic Frames (Goddard & Wierzbicka 2009, 2014, to appear; Wierzbicka 2009b, 2010; Ye 2010; Goddard 2014; Goddard, Wierzbicka & Wong, in press; Sibly 2010). The main business of this chapter is to develop an account of the alternation behaviors (specialized constructions) found with such verbs, with a particular focus on those that are included in the Leipzig Valency Patterns Project list. For reasons of space, we will not touch locomotion verbs, such as run, or the “bodily care” verbs, such as shave, wash, and dress. The verbs to be treated are: eat, pour, carry, dig and cut. Each belongs to a different subclass in the Levin (1993) classification, and, as we will see, they all have different Lexicosyntactic Frames.

Cliff Goddard

8

In §3.1–§3.5 we work through the selected verbs one at a time, roughly in order of the number and complexity of their alternations (specialized constructions). Except for eat, dealt with above, each subsection begins with the basic activity-in-progress sense, followed by a treatment of one or more alternations. The general picture is that the specialized constructions are quasi-derivational in nature, in the sense that the primary or semantically basic sense of the verb is embedded in a more elaborate configuration containing additional and often non-predictable semantic material2. Each such construction represents a kind of “word in construction” polysemy. 3.1 Eat (Plus: No eating or drinking and Have you eaten?)

The English verb eat is often said to exhibit an alternation termed the Omitted Object or Unspecified Object construction. Actually there are two such constructions, with semantic and grammatical differences between them. The simplest, and more grammatically restricted, construction is manifested also with read, and marginally, with drink. (1) No eating or drinking in class. (2) She was reading. The lack of an overt object implies disinterest in the nature or identity of what is being consumed. Whatever it is, it is normal, typical, expected. As noted by Wierzbicka (1988), one could say of a baby, for example: The baby is eating grass; but the same situation could not be described simply as ?The baby is eating. Likewise, a sentence like She was reading implies that the object was a book, magazine, newspaper or the like. One wouldn’t be likely to use such a sentence about someone reading a notice, a poster, or a product label. The meaning expressed by the omitted object can plausibly be represented as in the two explications below.

Explication [E]: Someone was eating2 (at this time). someone was eating [d] something at this time as people do in many places at many times

Explication [F]: Someone was reading2 (at this time).

someone was reading [d] something at this time as people do in many places at many times The second Omitted Object construction with eat is much more specific semantically, roughly equating with “eat a meal”. Unlike the general Omitted Object construction, it appears readily in the simple past and the perfect. (3) She was eating. (4) Cynthia ate alone. (5) Have you eaten? Explication [G] spells out the semantic content of the “eat a meal” construction (cf. the related expressions eat breakfast, eat lunch, eat dinner; compare: *eat snack). Essentially it represents the idea that someone who is eating (in the intended sense) is doing as people typically do at some times during the day when they spend time ‘eating some things of some kinds’. A feature of this explication is that the top-most component is not about ‘eating’, but rather about ‘doing something for some time’.

Explication [G]: Someone was eating3 (at this time). someone was doing something for some time at this time LEXICOSYNTACTIC FRAME it happened like this: this someone was eating [d] some things of some kinds as many people do

at some times during the day [m] MANNER

2 In broad terms this idea is reminiscent of the treatment of the locative and dative alternations in Rappaport and Levin (1988, 1998) and Pinker (1989); cf. Levin and Rappaport (2005: Ch 3).

Verb classes and valency alternations (NSM approach)

9

Interestingly, this construction is somewhat analogous to the expression ‘eat rice’ (meaning roughly, ‘eat a meal’)3 in many East Asian languages, even though from a formal point of view the role of the “object” is quite different. In English, there is no explicit object but this gap does not imply that anything at all is consumed. In Chinese, the object appears to be lexically rather specific (namely, rice), but it actually encompasses a broader range of food items4. 3.2 Pour (Plus: He poured me a drink, and Oil was pouring onto the street)

Examples of the basic activity-in-progress meaning of pour are shown below. (6) He was pouring it onto the ground. (7) She was pouring it out of (from) a jug. (8) She was pouring the water into a cup. The verb is explicated in [H]. Though water is depicted as the prototype material for pouring, both in the Prototypical Scenario and in characterising the “flowing” Manner of movement, nothing in the explication restricts the material to water: it could be any other liquid, or indeed, any other substance (such as grain or flour) that is capable of moving ‘like water [m] moves at many times’5. The distinctive property of water and similar substances is, of course, that one cannot manipulate them directly (Goddard 2010). Instead, one normally has the water or similar substance inside something, i.e. a vessel or container of some kind, and by manipulating this in a specific fashion (specifically, by raising part of it), one manages to control the flow of the water. The Prototypical Scenario for pour is more complex than that for eat. This is largely because the mental state appropriate for pouring is more complex than simple ‘wanting’. To pour something requires forming a conscious intention prior to the act, and, moreover, this intention has considerable internal complexity since it involves the idea not only of changing the location of the water or similar substance, but also the idea of manipulating the container for this purpose.

Explication [H]: Someone is pouring1 something (out of something) (at this time). someone is doing something to something for some time at this time because of this, something is happening to this something during this time

LEXICOSYNTACTIC FRAME

at many times when someone does this to something, it is like this: – this something is inside something else at this time – this something can move like water [m] can move – a short time before, this someone thought like this: “I want this something not to be inside this other thing after a short time, I want it to be

somewhere else because of this, I want this something to move for some time as I want it can be like this if I do something to this other thing for some time with my hands [m]”

PROTOTYPICAL SCENARIO

when someone does this to something, it happens like this: – this someone’s hands [m] move for a short time as this someone wants – because of this, some of this something moves at the same time as this someone wants – when it moves, it moves like water [m] moves at many times – because of this, after this, some of this something is not inside this other thing anymore

MANNER + EFFECT (HOW IT HAPPENS)

if this someone does this for some time, after this, all of this something can be somewhere not in the place where it was before

POTENTIAL OUTCOME

Although the quantity of material is not mentioned in [H], the verb pour implies a largish quantity (one couldn’t pour a single drop, for example). Hopefully this is sufficiently implied by the durational nature of the meaning as explicated.

3 In Cantonese, for example, the formula sik6 faan6 [eat rice] is what people generally say at the beginning of a meal. Sik6 zo2 faan6 mei6 aa3? ‘Have you eaten (rice) yet? [eat PFV rice/meal not-yet PRT] is often used as a general greeting, and also to ask whether someone has eaten yet. 4 Needless to say, the construction explicated in [G] has no analogue with the verb drink, which instead has its own independent “drink an alcoholic drink” meaning, cf. a drinker, a drinking problem. 5 Interestingly, the nearest Russian translation equivalent lit’ is restricted to liquids, while for grain or flour the appropriate verb is sypat’. Presumably the explication for Russian lit’ would not include the component ‘this something can move like water [m] can move’, but rather a component such as ‘this something is something like water [m]’.

Cliff Goddard

10

One of the interesting, and highly English-specific, constructions involving pour is the “production” meaning it has when used as “verb of preparation” (cf. Levin 1993). Examples like these usually have a Benefactive argument as well, and when they don’t (as in the last two examples below), a potential beneficiary is arguably implied. (9) He poured me a drink. (10) Pour me a glass of wine (beer, etc). (11) She taught me how to pour the perfect cup of tea. (12) I got a job pouring drinks in a cheap beachside bar. The presuppositions behind this usage are complex. They include the idea that we are talking about something of a kind that people normally drink in a certain way, which includes the liquid in question being in a glass or a cup. The actor’s intention is directed towards achieving a situation in which someone could drink in this way, by pouring some liquid into a glass. The potential to consume the drink can be seen as ‘good for this someone’, hence the commonly associated benefactive. The components in the Lexicosyntactic Frame and Outcome are presumably identical or close to those in semantically similar sentences such as: He made (me) a sandwich; He baked (me) a cake; She knitted (me) a sweater. Roughly, they involve producing something of one kind by some actions or activities. It seems likely that the “pour a drink” construction is modeled on these plainer verbs. The second line of the Lexicosyntactic Frame for pour2 is shared by the verb make (in the relevant sense) and very similar components can be attributed to carve and to build (Goddard & Wierzbicka in press). Explication [I] shows that there is a surprising amount of semantic detail – a kind of social scenario – implicit in the “pour a drink” construction.

Explication [I]: Someone poured2 (me) a drink. someone did something somewhere at this time because this someone wanted there to be something of one kind in this place at this time

LEXICOSYNTACTIC FRAME

because of this, after this, there was something of this kind in this place because of this, someone in this place could drink [m] this something as people drink [m]

something like this at many times

OUTCOME

it happened like this: – a short time before this something was inside something in this place – at this time this someone thought about this something like this: “it will be good for someone in this place if this someone can drink [m] some of this something

as people drink [m] something like this at many times because of this, I want some of this something to be inside something else (e.g. a glass) in

this place” – because of this, after this, this someone was pouring [d] this something for some time

MANNER (HOW IT HAPPENED)

On many treatments of alternation phenomena, sentences like the following would be counted as examples of an alternation involving pour (for example, as a kind of Ambitransitive or Anticausative alternation). A sample sentence is explicated in [J]. (13) Oil was pouring onto the street. (14) Smoke was pouring out of building.

Explication [J]: Oil was pouring3 onto the street. something was happening in this place for some time LEXICOSYNTACTIC FRAME it was happening like this: – much oil was moving in this place – before this, this oil was inside something – after this, it was somewhere else (on the street) – if people could see this oil when it was moving, they could think like this: “water [m] can move like this when someone is pouring [d] it”

HOW IT HAPPENS

Verb classes and valency alternations (NSM approach)

11

Another use of pour occurs in the expression pour (with rain). We will not explicate this in detail here6. Essentially, however, It was pouring means that it was raining (in a place) in a particular fashion which involves a great deal of water (very much water) and that this water was falling in a way which is not typical of normal rain, i.e. not in easily perceivable drops. The key component is much like the final component in [J], namely: ‘when people see this, they can think like this: “water [m] can move like this when someone is pouring [d] it”’. 3.3. Dig (Plus: He dug a hole and They were digging for gold)

The verb dig has two distinctive properties: first, it is inherently locational in the sense that it always implies action in a place and a concurrent effect in that place; second and more specifically, it implies an effect on the ground in that place. (In some languages, the close connection between ‘digging’ and the ground is manifested grammatically, in the fact that the noun for ‘ground’ appears as the default grammatical object of the verb; as for example, with the Russian imperfective kopat’ (inf.) which takes zemlju (acc.) ‘ground’ as default object.) The basic activity-in-progress meaning is explicated below.

Explication [K]: Someone is digging1 somewhere. someone is doing something somewhere for some time because of this, something is happening in this place during this time this someone is doing it with something

LEXICOSYNTACTIC FRAME

at many times when someone does this in a place, it is like this: – a short time before, this someone thought like this: “I want some of the ground [m] in this place to be after this time not in the place where it is now”

PROTOTYPICAL SCENARIO

when someone does this in a place, the same thing happens many times it happens like this: – this someone holds [m] something for some time – during this time, this someone does something to this something with parts of this someone’s body – because of this, part of this thing is inside the ground [m] in this place for a short time – during this time, this thing moves as this someone wants – because of this, something happens to the ground [m] in this place at this time – because of this, after this, some of the ground [m] is not in the place where it was before

MANNER + EFFECT (HOW IT HAPPENS)

if this someone does this for some time, after this, much of the ground [m] in this place is not in the place where it was before

POTENTIAL OUTCOME

There are two main specialized constructions involving dig, exemplified in the examples that follow. (15) a. He dug a hole (in the ground). b. They dug a well (tunnel, grave, trench). (16) He was digging for potatoes (gold, water, buried treasure). Digging a hole (well, tunnel, grave, trench, etc.) has a meaning structure much like that of “verbs of production/creation”, such as make and build. The nouns that can function as grammatical object are of course those denoting things that can be created by human activity upon or under the ground. Two slightly different explications are required. [L1] applies to the expression hole (in the ground). In [L1] the ‘digging’ is literal, so to speak, i.e. ‘dig [m]’ appears in the second part of the explication by itself. [L2] applies for other nouns, such as well, tunnel, grave, trench, etc. The final section of this explication makes use of a component type that we have not seen before. It can be termed an “analogy of action” component: ... this someone was doing something in this place for some time like someone does when someone is digging [m] in a place for some time 6 Presumably, one would want to treat to pour (with rain) as a figurative extension of pour, i.e. as a polysemic meaning, rather than as a valency alternation. One of themes of this chapter, however, is that from a semantic point of view the dividing lines between lexical polysemy, derivation and alternation are not strict (cf. Apresjan 1973).

Cliff Goddard

12

That is, when describing the manner in which someone digs a well (tunnel, grave, etc.), the activity is not depicted as ordinary digging, but as like ordinary digging. This is because the mechanics of “well digging” (tunnel digging, grave digging, etc.) do not necessarily, or even typically, exactly match ordinary digging, as previously explicated in [K]. “Analogy of action” components provide a mechanism for meaning extension.

Explication [L1]: Someone X dug2 a hole (in the ground) in this place. someone X did something in this place at this time because this someone wanted there to be a hole in the ground [m] in this place

LEXICOSYNTACTIC FRAME

because of this, after this, there was a hole in the ground [m] in this place OUTCOME it happened like this: this someone was digging [d] in this place for some time

HOW IT HAPPENED

Explication [L2]: Someone X dug2 a well (tunnel, grave, etc.) in this place.

someone X did something in this place at this time because this someone wanted there to be well (tunnel, grave, etc.) in this place

LEXICOSYNTACTIC FRAME

because of this, after this, there was a well (tunnel, grave, etc.) in this place OUTCOME it happened like this: this someone was doing something in this place for some time like someone does when

someone is digging [m] in a place for some time

HOW IT HAPPENED

Note that this analysis implies (and assumes) that the word hole is polysemous, with one of its meanings being specifically ‘hole in the ground’. The point is that a hole as such (unlike a well, tunnel, or grave), does not necessarily have to be in the ground. Many languages have different lexical forms for similarly different meanings, e.g. Yankunytjatjara piti ‘hole in ground’, ala ‘hole in something, opening’, cf. Spanish hoyo vs. agujero, Russian jama vs. dyra. Coming now to the “digging for” construction, it presupposes knowledge that things of certain kinds that people want to lay their hands on can be found in or under the ground. It seems that an element of uncertainty is part of it; cf. look for, search for. For example, if I have buried some treasure in the back yard for safekeeping and then go to dig it up, I will not be digging for treasure. It seems likely that the “digging for” construction is modeled on expressions such as look for and search for.

Explication [M]: Someone was digging3 for potatoes (water, gold, etc.) in this place. someone was doing something somewhere for some time because this someone wanted to do something with this something

LEXICOSYNTACTIC FRAME

it happened like this: – some time before, this someone thought like this about this place: “maybe there is something good of one kind (potatoes, water, gold) below the ground [m] in

this place I want to know if there is something like this below the ground [m] in this place, I want it not be below the

ground [m] anymore I want to do something with it” – because of this, this someone wanted to do something to the ground [m] in this place – because of this, this someone was doing something in this place for some time like someone

does when someone is digging [m] in a place for some time

HOW IT HAPPENED

The verb dig can also participate in “path” constructions, such as the following, but for reasons of space, these will not be explicated here. (17) They dug through the top layer of shale. (18) They dug their way out of the wreckage [Way-construction] 3.4 Carry (Plus: He carried a knife)

The activity-in-progress meaning of carry can be explicated as in [N] below. We can quickly note a couple of details. First, the second component in the Lexicosyntactic Frame (‘because of this, this something is not in one place during this time, it is in many places at many times’) conveys the idea of

Verb classes and valency alternations (NSM approach)

13

on-going displacement. Second, components in the Prototypical Scenario and Potential Outcome (‘if this someone does this for some time, after this, this something can be somewhere else as this someone wanted before’) convey the idea that the thing being carried can end up in a particular intended destination. Third, components in the Prototypical Scenario and Manner sections capture the fact that carrying involves avoiding contact with the ground. This is necessary to distinguish carrying from dragging, for example. Fourth, the explication provides that contact with part of the agent’s body is an important part of the meaning, but without specifying any particular part. This is compatible with the possibility of adding phrases such as in his arms, on his back, on her head, etc., to identify the relevant body-part. (Many languages lack any verb with a comparably broad meaning, but instead have more specialized verbs distinguished by body-part and associated manner.)

Explication [N]: Someone is carrying1 something (e.g. the boxes, a baby). someone is doing something to something for some time because of this, this something is not in one place during this time, it is in many places at many

times this someone is doing it with some parts of the body

LEXICOSYNTACTIC FRAME

at many times when someone does this to something, it is like this: – some time before, this something was in the place where this someone was – this someone thought like this at this time: “I don’t want this something to be in this place after this I want it to be somewhere else after some time because of this, I want to do something to it for some time after this I don’t want it to touch the ground [m] during this time”

PROTOTYPICAL SCENARIO

when someone does this to something, it happens like this: – for some time parts of this someone’s body touch parts of this thing as this someone wants – at the same time this someone does something with some other parts of the body – because of this, this someone’s body is not in one place during this time, it is in many places at

many times

MANNER + EFFECT (HOW IT HAPPENS)

if this someone does this for some time, after this, this something can be somewhere else as this someone wanted before

POTENTIAL OUTCOME

With carry, the main alternative construction of interest is illustrated in the sentences below. This can be termed the Quasi-Benefactive Accompaniment alternation. (19) She carried a notebook/lip balm (with her) at all times. (20) Max carries a knife (with him) everywhere he goes. In this meaning, which I designate carry2, the verb can optionally occur with a prepositional phrase of the form with + Pronoun, where the pronoun indexes the agent. Semantically and grammatically, the construction is related to have something with (one), e.g. She always has her notebook with her. As noted by Levin (1993: 104), the prepositional phrase has the unusual property that the pronoun cannot appear in reflexive form, despite it being co-referential with the subject, cf. *She carried it with herself.7 Unlike carry1, the verb carry2 cannot take a goal expression. (21) ?She carried the lip balm (with her) downstairs. (22) ?He carried a knife (with him) downstairs. Another distinctive grammatical property of carry2 is that it does not have a past perfective. A sentence like He carried2 a weapon, for example, must have a past habitual interpretation. A notable distributional difference between carry1 and carry2 is that in the case of the latter, the object can be something small, or even very small, such as a pen or lipstick. The carry2 meaning is explicated in [O]. Notice that the Prototypical Scenario is quite different to that of carry1: essentially, the actor is envisaging being “mobile” for some time and considers it potentially beneficial to be able to do something with the item in question during this time; hence, he 7 Possibly the factor that licenses the non-reflexive pronoun is the implicit reference in explication [O] to the actor’s body, rather than to the actor him or herself. As Beth Levin (p.c.) has observed, similar non-reflexive sentences are found with verbs push and pull, e.g. She pushed the plate away from her. NSM explications for push and pull also involve reference to proximity to the actor’s body. The matter requires further research.

Cliff Goddard

14

or she wants to have the item ‘near my body at all times during this time’. This scenario is highly compatible with functional items such as weapons, writing instruments, mobile devices, and the like, but it is sufficiently broad to accommodate sentences such as She carried a Bible at all times and He carried her photo with him at all times. The Manner section shows a new type of analogy component expressing what can be termed “manner of effect”. when someone does this to something, something happens to this something like something happens to something when

someone is carrying [d] this something Rather than the agent’s action being likened to the primary sense of the verb, the analogy concerns the effect on the object: when someone carries2 something, the effect on this object is like what happens to something when someone carries1 something.

Explication [O]: Someone was carrying2 something Y (with him/her). someone was doing something to something for some time (at this time) because of this, this something was not in one place during this time, it was in many places

LEXICOSYNTACTIC FRAME

at many times when someone does this to something, it is like this: – a short time before this, this someone thought about this something like this: “for some time after this I won't be in one place, I will be in many places at many times it can be good if at some time during this time I can do something with this something because of this, I want this something to be near my body at all times during this time”

PROTOTYPICAL SCENARIO

when someone does this to something, something happens to this something like something happens to something when someone is carrying [d] this something

MANNER

The existence of carry2 in itself does not represent a valency alternation of carry. On the other hand, carry2 has a distinct valency alternation associated with it, namely, the option of including the with-phrase. 3.5 Cut (Plus: He cut his face while shaving, He cut his foot on a rock, and The glass cut his hand)

The lexical semantics of cut and chop, and their nearest equivalents in Polish and Japanese, have been analyzed in Goddard & Wierzbicka (2009). The present treatment generally follows that analysis with some adjustments and revisions. The activity-in-progress meaning can be explicated as in [P]. (23) She was cutting the bread. (24) He cut the paper with scissors.

Explication [P]: Someone is cutting1 something (e.g. bread, paper). someone is doing something to something for some time because of this, something is happening to this something during this time as this someone wants this someone is doing it with something else

LEXICOSYNTACTIC FRAME

at many times when someone does this to something, it is like this: – a short time before, this someone thought like this about this something:

“I don’t want this thing to be one thing anymore, I want it to be two things because of this, I want to do something to it for some time after this when I do this, I want something to happen to it all the time as I want”

PROTOTYPICAL SCENARIO

when someone does this to something, it happens like this: – this someone holds [m] part of something else with one hand [m] all the time – some parts of this other something are sharp [m] – this someoneʼs hand [m] moves for some time as this someone wants – because of this, the sharp [m] parts of this other thing touch this thing in some places as this

someone wants – because of this, something happens to this thing in these places as this someone wants – because of this, after this, part of this thing is not like it was before

MANNER + EFFECT (HOW IT HAPPENS)

if this someone does this for some time, after this, this thing can be two things as this someone wanted before

POTENTIAL OUTCOME

Explication [P] above will not work for uses of cut in which the object is something very thin, such as string, thread or ribbon, because cutting things like these does not involve a activity carried out over

Verb classes and valency alternations (NSM approach)

15

a period of time. On the contrary, the agent does something and the effect happens ‘in one moment’. Naturally therefore, when one cuts a string (thread, ribbon, etc.) there is no possibility of exercising sustained ongoing control. Examples of cut with objects like these strongly tend to occur in perfective contexts: (25) He finished off the last stitch, knotted it and cut the thread with her nail scissors. (26) He held the parcel on his knees and cut the string with a clasp knife. Uses like these can be explicated as in [Q], which is given in the past perfective frame.

Explication [Q]: He cut2 the string (thread, ribbon). he did something to the string at this time because of this, something happened to the string at this time it happened in one moment

LEXICOSYNTACTIC FRAME

because of this, after this, the string was not one thing anymore, it was two things OUTCOME it happened like this: – he did something with something sharp [m] – because of this, something happened to this string like something happens to something when someone is cutting [d] it

HOW IT HAPPENED

English cut can appear in several specialized constructions connected with the unintended effects of a sharp object on part of a person’s body. We will consider three of these, exemplified in the three sentences below. They can all be termed Accidental Body-Part Damage alternations. On account of these and similar alternations, Levin (1993) cross-listed cut both as a “Verb of Cutting” and as a “Hurt verb” – involving “damage to the body through a process that is not under control of the subject”. The events being described are usually very brief in duration (naturally enough) and consequently these constructions do not normally occur in the progressive. Note that the “damage to the body” events described by these sentences do not imply any actual or potential separation into two parts – another difference from canonical “cutting”. (27) He cut his face while shaving. (cf. *He was cutting his face while shaving.) (28) He cut his foot on a rock. (cf. *He was cutting his foot on a rock.) (29) The glass cut his hand. (cf. *The glass was cutting his hand.) From an English speaker’s point of view, these constructions shown in (27)–(29) may seem natural enough but for speakers of languages which do not have such constructions8, they can seem puzzling. What is their semantic rationale? The answer, evidently, lies in the involvement of something sharp, and more specifically, in the fact that contact with something sharp can damage the human body. The way in which damage occurs can be seen as like what happens to something when it is being cut. Both the explications below rely on a “manner of effect” component. The effect of the sharp object on the affected part of the body is characterized as follows: – (because of this,) something happens to this part of his/her body like something happens to something when someone is cutting [d] this something Explication [R] is for He cut his face while shaving. Comparable examples would include: She cut her hand grating the carrots, He cut his hand while slicing bagels, and the like. The Lexicosyntactic Frame and Outcome sections characterize the meaning, at a broad level, as doing something to part of one’s body which causes something unwanted to happen to it which has a resulting localized effect that lasts for some time. The scenario for How it Happened involves the actor doing something for some time with a sharp instrument, during which time a sharp part of this instrument touches part of 8 In Russian, for example, the ‘cut’ verb would normally require a prefix po- in sentences like (27)-(29). In Polish, one would use a verb like ‘injure’ for (27), rather than ‘cut’, while for the others, ‘cut’ could be used but only with a prefix meaning something like ‘through’ (Anna Wierzbicka p.c.). In Chinese, yet another pattern is evident. (27)-(29) could be rendered using the verbal expression gē-pò [cut-broken], either by itself for (27), or in passive or ba-constructions, with (28) and (29), respectively, but this is quite different from the normal ‘cut’ verb qie, which implies deliberate and controlled action (Li Jiongying p.c.; Zhengdao Ye p.c.).

Cliff Goddard

16

the actor’s body ‘not as this someone wanted’. This wording is compatible with a situation of shaving, as well as other kinds of activity (such as grating, slicing, etc.). The contact causes something to happen to the body-part in question, and, as just explained, the manner of effect is described as being like what happens to something when someone is cutting it. Finally, the event is characterized as momentary: ‘it happened in one moment’.

Explication [R]: He cut3 his face while shaving. he did something to part of his body (his face) at this time because of this, something happened to his face at this time he didn’t want this to happen it happened in one moment

LEXICOSYNTACTIC FRAME

because of this, after this, part of his face was for some time not like it was before OUTCOME it happened like this: – he was doing something for some time with something sharp [m] – at some time during this time, this sharp [m] thing touched part of his face not as he wanted – because of this, something happened to this part of his face like something happens to something when someone is cutting [d] this something

HOW IT HAPPENED

Not surprisingly, other verbs implying contact with something sharp can also appear in this construction; for example: I scratched my hand while coiling the wire. There is another cut-construction, semantically quite close to that explicated above, exemplified by sentences such as: He cut himself on the face (while shaving). It differs grammatically in employing the reflexive pronoun and an on-phrase indicating the “locus of contact”. This construction will not be discussed here, cf. Goddard & Wierzbicka (2009: 69–70). Explication [S] is for He cut his foot on a sharp rock. This sentence could be used about a situation, for example, in which someone was climbing around bare-footed on some rocks. Other comparable examples would be: He cut his hand on the barbed wire, I cut my hand on a piece of broken glass. The construction implies that the actor was doing something in a place where there was something sharp and that during this time part of his or her body came into accidental contact with the sharp thing, with a resulting “cutting-like” effect. As with the previous explication, the final component of the Lexicosyntactic Frame states that this happened in one moment.

Explication [S]: He cut4 his foot on a rock. something happened to part of his body (his foot) at this time it happened because he was doing something somewhere at this time he didn’t want this to happen it happened in one moment

LEXICOSYNTACTIC FRAME

because of this, after this, this part of his body was for some time not like it was before OUTCOME it happened like this: – there was something sharp [m] in the place where he was – when he was doing something in this place, part of his body touched this sharp [m] thing not as

he wanted – because of this, something happened to this part of his body like something happens to something when someone is cutting [d] this something

HOW IT HAPPENED

As before, other verbs implying contact with something sharp can also appear in this construction; for example: I nicked my arm on the edge, I pricked my finger on a thorn. Verbs like burn, hit, and scrape can participate in the same construction, as shown in the examples below. In place of ‘something sharp [m]’, such examples presumably imply, respectively, the presence of ‘something very hard [m]’, ‘something very hot [m]’, ‘and ‘something very rough [m]’. Items with these properties all have the potential to cause harm to the human body upon contact. (30) He hit his head on the roof as he stood up. (31) My four year old son burned his hand on the stove. (32) I scraped my finger on the asphalt. The following examples show the third and final Accidental Body-Part Damage construction with cut to be considered here. The sharp object appears as the grammatical subject.

Verb classes and valency alternations (NSM approach)

17

(33) The glass cut her hand. (34) As she pushed through the dense bush, the razor-sharp leaves cut her face terribly. (35) I was clearing the pieces away and one of them cut my hand. Such sentences are rather vague as to the nature of the causing situation, i.e. how it comes about that the sharp thing denoted by the subject comes into contact with part of the patient’s body. It is notable, however, that the construction does not extend across the same range as the one explicated in [S]. (36) *The roof hit his head. (37) *The stove burned my son’s hand. (38) *The asphalt scraped my finger. Comparing these unacceptable sentences with more acceptable ones, such as He hit his head on the roof and He burned his hand on the stove, suggests that the crucial requirement for the damaging item to assume the role of grammatical subject is movement. Some sharp items, especially small ones like pieces of glass, thorns, etc., are capable of moving – not of their own volition, of course, but as a result of something that happens in the situation. In many cases, the movement is caused by the person’s own actions, but this is not necessarily the case, as we can see from sentences such as the following. (39) There was an explosion and a piece of flying glass cut her face. (40) Something exploded in the fire and a flying spark burned her face. Explication [T] requires that something sharp in the immediate vicinity of the actor moves very briefly and in so doing comes into contact with part of someone’s body9.

Explication [T]: The glass cut5 her hand. something moved in the place where she was at this time because of this, something happened to part of her body (her hand) at this time she didn’t want this to happen it happened in one moment

LEXICOSYNTACTIC FRAME

because of this, after this, this part of her body was for some time not like it was before OUTCOME it happened like this: – there was something sharp [m] in the place where she was at this time – this sharp [m] thing moved for a very short time at some time – during this time it touched part of her body – because of this, something happened to this part of her body like something happens to something when someone is cutting [d] this something

HOW IT HAPPENED

4. Review and discussion Levin and Rappaport Hovav (2005) were right when they asserted that progress towards improved understanding of valency classes and alteration phenomena will depend on detailed lexical semantic analysis. Two questions they left hanging were: How detailed does the analysis have to be? And what is the appropriate methodology for getting to the required level of detail? On the basis of the present study, we can conclude that to get a clear view of valency phenomena what is needed is explication right down to the level of semantic primes and molecules; and (relatedly) that the NSM approach provides a suitable methodology. The task is a formidable one, but in my view the usual strategy of using complex-but-approximate labels for putative “syntactically relevant” meanings has largely exhausted its usefulness10. It is time to take the plunge and get to work with detailed lexical semantic 9 In some sentences, such as (34), and the similar The bush scratched my face as I rushed blindly down the path, the actor is also moving and it is this that initiates the causal chain that leads to him or her getting cut. Even so, the construction under discussion seems to imply some movement by the sharp thing, e.g. the razor-sharp leaves, thorns, bush, etc. The construction would not be acceptable about an immobile object, such as a rough wall, for example; *The wall scratched my face as I rushed blindly down the narrow path. 10 The present study raises certain problems for Levin and Rappaport Hovav’s proposals about “Manner/Result complementarity” (cf. Levin, this volume; Rappaport Hovav & Levin 2010). For example, it is clear from the present study

Cliff Goddard

18

analysis. Aside from clarifying the nature of verb classes and alternation phenomena (specialized constructions), detailed lexical semantic analysis is necessary in order to settle confounding issues concerning lexical polysemy. So far as the semantic character of alternations (specialized constructions) is concerned, it has emerged from the present study that syntactic alternations express rich configurations of meaning that combine the lexical semantics of the primary sense of the verb with additional meanings. This makes construction semantics seem analogous to derivation. As Dowty (2000) says (he is writing about the Locative-subject construction, but his points apply more broadly):

[C]ontrary to the usual view …, good reasons can be given to view it as a lexical derivation analogous to rules of WORD FORMATION on the one hand, and to processes of LEXICAL SEMANTIC EXTENSION … and METAPHOR on the other. (Dowty 2000: 121; emphasis in original)

The account developed in the present study confirms this view and, moreover, shows the mechanics of this: the derivation-like nature of the specialized constructions is visible in the involvement of the primary meaning as a [d] element in the composite meaning. The idea has also emerged that specialized constructions for a given verb are often modeled on the semantics and constructions of other verbs. For example, we suggested that the “pour a drink” construction is partly modeled on expressions such as make a sandwich and knit a sweater, and that the “digging for gold” construction is partly modeled on expressions such as look for gold and search for water. This is broadly consistent with the proposal by some cognitive linguists and scholars in construction grammar that many constructions can be seen as lexicosyntactic “blends” (Goldberg 1995, 2010; Fauconnier & Turner 1996; Dancygier 2009). Much remains to be done, however, before these processes can be regarded as well understood. On account of its microsemantic nature, the present study has been highly delimited in many ways, restricted to one subdomain in a single language, i.e. English, which uses “phrasal constructions” (lexicosyntactic blending) as its main mode for expressing specialized meaning alternations. As the data being marshalled by the Valency Classes Project shows, similar effects can be achieved in other languages by very different morphosyntactic means, such as derivational morphology, applicative morphology, or serial verb constructions. It is reasonable to expect that the semantics of these processes is broadly similar to those that we see in English, but “broadly similar” does not mean identical. As richer and more comparable cross-linguistic data comes to hand, the challenge will be to apply detailed semantic analysis to this data using standardized and consistent methods.

that the verb cut contains detailed manner components, which runs counter Levin & Rappaport Hovav’s (to appear) arguments that cut is “basically a result verb” and that no manner components are lexicalized as part of its meaning. No doubt their key insight could be re-formulated, for example, in terms of “manner orientation” vs. “result orientation”. I remain doubtful, however, about the extent to which simple bifurcations can illuminate the complexities of alternation phenomena. It also seems to me that generalizations at the level of the lexeme are necessarily perilous on account of lexical polysemy and because alternation phenomena are intertwined with grammatical aspect, as well as lexical aspect.

Verb classes and valency alternations (NSM approach)

19

Appendix 1. Semantic primes (English exponents), grouped into related categories (February 2013)

I~ME, YOU, SOMEONE, SOMETHING~THING, PEOPLE, BODY substantives KIND, PART relational substantives THIS, THE SAME, OTHER~ELSE determiners

ONE, TWO, SOME, ALL, MUCH~MANY, LITTLE~FEW quantifiers GOOD, BAD evaluators

BIG, SMALL descriptors KNOW, THINK, WANT, DON’T WANT, FEEL, SEE, HEAR mental predicates SAY, WORDS, TRUE speech

DO, HAPPEN, MOVE actions, events, movement, contact BE (SOMEWHERE), THERE IS, BE (SOMEONE/SOMETHING), (BE) MINE

location, existence, specification, possession

LIVE, DIE life and death WHEN~TIME, NOW, BEFORE, AFTER, A LONG TIME, A SHORT TIME, FOR SOME TIME, MOMENT

time

WHERE~PLACE, HERE, ABOVE, BELOW, FAR, NEAR, SIDE, INSIDE, TOUCH space

NOT, MAYBE, CAN, BECAUSE, IF logical concepts VERY, MORE intensifier, augmentor LIKE~AS~WAY similarity

Notes: • Primes exist as the meanings of lexical units (not at the level of lexemes) • Exponents of primes may be words, bound morphemes, or phrasemes • They can be formally complex • They can have combinatorial variants or “allolexes” (indicated with ~) • Each prime has well-specified syntactic (combinatorial) properties.

Acknowledgements I am very grateful to the organisers and to my fellow participants in the Leipzig Valency Patterns Project both for their direct and indirect contributions. The editors of this volume and reviewers made valuable suggestions and observations. The explications in this chapter have been discussed extensively with Anna Wierzbicka. References Apresjan, Juri D. 1973. Regular polysemy. Linguistics 124: 5-32. Barrios Rodriguez, María Auxiliadora & Goddard, Cliff 2013. “Degrad verbs” in Spanish and English: Collocations, Lexical Functions and contrastive NSM

semantic analysis. Functions of Language. 29(2), 219-249. Bromhead, Helen 2009. The Reign of Truth and Faith: Epistemic expressions in 16th and 17th century English. Berlin: Mouton

de Gruyter. 2011. Ethnogeographical categories in English and Pitjantjatjara/Yankunytjatjara. Language Sciences 33(1),

58-75. Comrie, Bernard 1993. Aspect. Cambridge: Cambridge University Press. Dancygier, Barbara 2009. Genitives and proper names in construction blends. In: Evans, Vyvyan & Pourcel, Stéphanie (eds.),

New Directions in Cognitive Linguistics, 161-184. Amsterdam: John Benjamins. Dowty, David 2002. ‘The Garden Swarms with Bees’ and the Fallacy of ‘Argument Alternation’. In: Ravin, Yael &

Leacock, Claudia (eds.), Polysemy: Theoretical and Computational Approaches, 111–128. Oxford: Oxford University Press.

Cliff Goddard

20

Fauconnier, Gilles and Turner, Mark 1996. Blending as a central process of grammar. In: Goldberg, Adele (ed.), Conceptual Structure, Discourse

and Language, 113–130. Stanford: CSLI. Gladkova, Anna 2010. Russkaja kul’turnaja semantika: ėmocii, cennosti, žiznennye ustanovki [Russian cultural semantics:

Emotions, values, attitudes.] Moscow: Languages of Slavic Cultures. Goddard, Cliff 2010. Semantic molecules and semantic complexity (with special reference to “environmental” molecules).

Review of Cognitive Linguistics 8(1): 123-155. 2011. Semantic Analysis [Revised 2nd edition]. Oxford: Oxford University Press. 2012. Semantic primes, semantic molecules, semantic templates: Key concepts in the NSM approach to

lexical typology. Linguistics 50(3): 711–743. Goddard, Cliff (ed.) 2008. Cross-Linguistic Semantics. Amsterdam: John Benjamins. Goddard, Cliff & Wierzbicka, Anna 2007. NSM analyses of the semantics of physical qualities: sweet, hot, hard, heavy, rough, sharp in cross-

linguistic perspective. Studies in Language 31(4): 675-800. 2009. Contrastive semantics of physical activity verbs: ‘cutting’ and ‘chopping’ in English, Polish, and

Japanese. Language Sciences 31: 60-96. 2014. Words and Meanings. Lexical Semantics Across Domains, Languages and Cultures. Oxford: Oxford

University Press. In press. Explicating the English lexicon of “doing” and “happening”. Functions of Language. Goddard, Cliff & Wierzbicka, Anna (eds.) 2002. Meaning and Universal Grammar – Theory and Empirical Findings. Vols I and II. Amsterdam: John

Benjamins. Goddard, Cliff, Wierzbicka, Anna & Wong, Jock In press. “Walking” and “running” in English and German: The cross-linguistic semantics of verbs of human

locomotion. Review of Cognitive Linguistics. Goldberg, Adele 1995. Constructions: A Construction Grammar approach to argument-structure constructions. Chicago:

University of Chicago Press. 2010. Verbs, constructions and semantic frames. In: Rappaport Hovav, Malka, Doran, Edit & Sichel, Ivy

(eds.), Lexical Semantics, Syntax, and Event Structure, 39–58. Oxford: Oxford University Press. Habib, Sandy 2011. Contrastive lexical conceptual analysis of folk religious concepts in English, Arabic and Hebrew:

NSM approach. PhD thesis, University of New England. Levin, Beth 1993. English Verb Classes and Alternations. University of Chicago Press: Chicago. this volume. Verb classes within and across languages. Levin, Beth & Rappaport Hovav, Malka 2005. Argument Realization. Cambridge: Cambridge University Press. to appear. Lexicalized meaning and Manner/Result complementarity. Ms.

[www.stanford.edu/~bclevin/barcel11rev.pdf, accessed 17/01/2013]. Levisen, Carsten 2012. Cultural Semantics and Social Cognition: A case study on the Danish universe of meaning. Berlin:

Mouton de Gruyter. Majid, Asifa & Bowerman, Melissa (eds.) 2007. “Cutting and breaking events”: a cross-linguistic perspective. Special Issue of Cognitive Linguistics

18(2). Peeters, Bert (ed.) 2006. Semantic Primes and universal grammar: Empirical evidence from the Romance languages.

Amsterdam: John Benjamins. Pinker, Steven 1989. Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, MA: MIT Press. Rappaport Hovav, Malka & Levin, Beth 1988. What to do with theta-roles. In: Wilkins, Wendy (ed.), Syntax and Semantics 21: Thematic Relations,

7–36. New York: Academic Press. 1998. Building verb meanings. In: Butt, M. & Geuder, W. (eds.), The Projection of Arguments: Lexical and

Compositional Factors, 97–134. Stanford, CA: CSLI Publications.

Verb classes and valency alternations (NSM approach)

21

2010. Reflections on Manner/Result Complementarity. In: Rappaport Hovav, Malka, Doran, Edit & Sichel, Ivy (eds.), Lexical Semantics, Syntax, and Event Structure, 21–38. Oxford: Oxford University Press.

Sibly, Anne 2010. Harry slapped Hugo, Tracey smacked Ritchie: The semantics of slap and smack. Australian Journal of

Linguistics 30(3): 323–348. Vanhatalo, Ulla, Heli Tissari & Anna Idström. 2014. Re-visiting the universality of Natural Semantic

Metalanguage: A view through Finnish. SKY Journal of Linguistics 27. 67–94. Wierzbicka, Anna 1988. The Semantics of Grammar. Amsterdam: John Benjamins. 1996. Semantics: Primes and Universals. New York: Oxford University Press. 2002. Semantic Primes and Universal Grammar in Polish. In: Goddard, Cliff & Wierzbicka, Anna (eds.),

Meaning and Universal Grammar – Theory and Empirical Findings. Vol I, 65–146. Amsterdam: John Benjamins.

2008. Case in NSM: A reanalysis of the Polish Dative. In: Malchukov, Andrej & Spencer, Andrew (eds.), The Oxford Handbook of Case, 151–169. Oxford: Oxford University Press.

2009a. The theory of the mental lexicon. HSK (Handbücher zur Sprach und Kommunikationswissenschaft. Handbook of Linguistics and Communication Science) “Slavic Languages”, 848–863. Berlin: Mouton de Gruyter.

2009b. All people eat and drink: Does this mean that ‘eat’ and ‘drink’ are universal human concepts? In: Newman, John (ed.), The Linguistics of Eating and Drinking, 65–89. Amsterdam: John Benjamins.

Wong, Jock 2005. “Why you so Singlish one?” A semantic and cultural interpretation of the Singapore English particle

one. Language in Society 34(2): 239–275. 2010. The triple articulation of language. Journal of Pragmatics 42(11): 2932–2944. 2014. The Culture of Singapore English. Cambridge: Cambridge University Press. Ye, Zhengdao 2010. Eating and drinking in Mandarin and Shanghainese: A lexical-conceptual analysis. In: Christensen, E.

Schier, E. & Sutton, J. (eds.), ASCS09: Proceedings of the 9th Conference of the Australasian Society for Cognitive Science, 375–83. Sydney: Macquarie Centre for Cognitive Science.

(ed.) In press. The Semantics of Nouns. Oxford: Oxford University Press.

Recommended