Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống
1
/ 59 trang
THÔNG TIN TÀI LIỆU
Thông tin cơ bản
Định dạng
Số trang
59
Dung lượng
1,74 MB
Nội dung
experiencer. On the basis of Chapter 9, we can formulate the further gener- alization that the relations borne by circumstantials are limited strictly to the locative subdomain. Figure 13.1 also indicates, by the arrows, the dependence of the second- order feature {goal} on the presence of a source. And the figure represents absolutive as lying in a domain that includes the two subdomains, but also as standing outside them: thus it may participate autonomously from either domain in processes or states. The two major zones occupied by the other semantic relations in each of which absolutive may be included are the domain of action and that of location: absolutive may be acted upon or located. The prototypical other participants in these subdomains are, respect- ively, the (prototypically second-person) volitional agentive which presents the most palpable causal source in the representation of a scene, on the one hand, and, on the other, the concrete spatial location that forms the percep- tually most accessible ground in a representation. These prototypes define the endpoints of (respectively) ‘animacy hierarchies’ (for example Silverstein 1976; DeLancey 1981) and a dimension of relative concreteness and dimen- sional differentiation. The experiencer, {{source,locative}}, unites the two subdomains, and such a participant is prototypically first-person. On the different prototypicalities of first and second person, Wierzbicka (1981: 46) comments: The speaker is more interested in what other people are doing to him than in what he is doing to other people; he is more sensitive to the ways in which other people’s actions affect him than to the ways in which his actions affect other people. The speaker regards himself as the quintessential ‘victim’ or the quintessential experi- encer. Whatever the status of this, the distribution emerges from various typological observations concerning ‘animacy hierarchies’. Once more, however, as concerns the criteria we have been looking at, they have a reduced significance in the context of a restricted set of semantic relations. More important for lexical structure as a whole is the fact that the constraints on it are ‘local’ requirements associated with the semantic- relational categories and their arguments: they are imposed by valencies and modifications, signalled by ‘/’ and ‘\’, respectively. This is indeed a property shared with the syntax; but syntax differs in that it builds structures relating different lexical items and it imposes linearity on them, at least partially. Epilogue: Case, Notionalism, Creativity, and the Lexicon 403 13.2.4 Lexical structure and morphology Given these differences, syntactic structure, as well as being built on the basis of valency and modification, also involves interaction among secondary features of different items, in particular (in traditional terms) agreement and rection; and these are expressed morphologically. Rection involves determination of features by a particular primary category. The classic traditional example in English is the determination of pronouns as accusative when governed by a preposition or verb—though, as we have seen, the situation is more complex than that. This represents an extreme routinization of the interaction between predicator, complex functor, and morphological case that we find in Latin, or Finnish (as described in Chapter 8). Simple agreement involves a matching of the features belonging to particular secondary categories—in procedural terms, copying of features from one category to another. With verb agreement, for instance, this may rather reflect incorporation of an argument. More complex in terms of morphosyntax are the Basque systems of agree- ment, where, for instance, in the predominant ‘analytic’ (rather than ‘syn- thetic’) verbal construction, the categories associated with the incorporated arguments of a verb are expressed on the governing operative: (7.34) a. Aitak ogia jaten du father: SG.DEF.ERG bread:SG.DEF.ABS eating 3SG.ABS:3SG.ERG (‘(My) father is eating the bread’) b. Amak aita maitatzen du mother: SG.DEF.ERG father: SG.DEF.ABS loving 3SG.ABS:3SG.ERG (‘(My) mother loves (my) father’) I suggestthat,aswith the apparent ‘complements’ofcomplex (deverbal) nouns, the (optional) nominals here are in apposition with incorporated arguments. In (21a) there are three agreeing incorporated arguments, but all expressed on theoperative rather thanon the lexicalverb towhich theybearsemantic relations: (21) a. Eman diozkat given 3 PL.ABS:3SG.DAT:1SGERG (‘gave her/him them’) b. {P} ………. {P;N} {erg} {{erg,loc}} {abs} | | | {N{1 SG}} {N{3 PL}}{N{3SG}} 404 Modern Grammars of Case We can indicate the incorporated arguments in (21), schematically, as in (21b), where the co-dependent incorporated functors are not linearized. These dependents are replicated on the operative, roughly indicated by the double-headed arrow; and they are expressed there. So we also have again non-expression of incorporated arguments in the case of the verb—though they are expressed elsewhere, on the operative. Morphology thus expresses secondary features ‘inherited’ by agreement, as well as inherent features such as gender on nouns. Expression may involve in- ternal modifications (including the drastic form of this known as suppletion) or affixation. In the latter case, of course, linearity is imposed by the morphology. Determination of linearity is a property shared with syntax. But, as we have noted, lexical structure in general has a capacity not permitted to the syntax: it can change syntactic categories, or, rather, subjoin the category of a base to a distinct derived category. This involves both primary and secondary features. Anderson (2003) offers the notation in (22 ) to represent instances of this capacity and the role played by affixes: (22) a. {P:N} b. {N;P} | | {N;P} {\{N;P }\\{ P:N}} {N;P} {\{N;P}\\ {N;P{abs}} : : : : : : : : beauty ful man hood Ful is an affix that seeks to modify a noun (‘\’) which is converted (‘\\’) to an adjective; hood is associated with change of subclass (to what might be glossed as ‘abs(tract)’). But with conversions, such as those in (23), involving the derived noun walk, the derived verb table and the derived count noun (a) beauty, no such affixation is deployed: (23) a. {N;P} b. {P;N} c. {N;P{count}} | | | {P;N} {{loc{goal}}} {N;P{abs}} : | : : {N;P} : walk : beauty : table The lexical entries in (23) are nevertheless categorially complex (though the representations are obviously incomplete as they stand), and show a relation- ship to the relevant base. This is no more than to say that we have both overt derivation and conversion. Epilogue: Case, Notionalism, Creativity, and the Lexicon 405 One proposed restriction on lexical structure that can’t be supported is that made concerning conversions by Beard (1998: 62), who claims: For every conversion to dry, to wet, to empty, we find an equal number of affixed derivates with the same relation: to shorten, to normalize, to domesticate. Moreover, precisely those stems which affix are precluded from conversion (to * short, * normal, * domestic), and precisely those which convert are precluded from affixation: to * endry, * wetten, * emptify. However, Sanders (1988 ) has shown that the existence of equivalent overt marking—what he calls the ‘overt analogue criterion’—cannot be maintained in general, given the relative paucity of overt derivational morphology in various languages (see too Colman and Anderson 2004). And such an as- sumption underestimates the productivity of conversion in English, for instance. We still have a long way to go in the study of such expression- based constraints as Beard aims at. Conversions typically involve ‘changes’ in category, and absorption, which are directly responsible for other changes (in valency, for instance). A number of affixations add a particular component of meaning that doesn’t entirely follow from the category or valency change, as with the -able formations mentioned in §4.2.2, and illustrated by the (b) examples in (4.26) and (4.27): (4.26) a. The meeting day can be changed/varied b. The meeting day is changeable/variable (4.27) a. The weather can change/vary b. The weather is changeable/variable Here the suffix adds a very particular component of interpretation, one specifically associated with the presence of the suffix. We might represent this, crudely, as in (24), where ‘pot(ential)’ abbreviates whatever (modal/ aspectual) specification(s) might be appropriate: (24) {P:N} | {P;N} {{pot}\{P;N}\\{P:N}} | : <{{erg}}> : | : <{N}> : : : : : vary able 406 Modern Grammars of Case Any ergative argument of the base verb is incorporated. Conversions typically involve simple change in (primary or secondary) category and its conse- quences. There is thus also a difference in the character of the typical derivations involving conversions and affixation. Notice too that, despite being associated with a difference in syntactic category, the affixes in (22) are not syntactic heads; indeed, they are not heads of anything (pace Williams 1981). They are not independent syntactic elements, and as morphological (syntactic-category-free) elements, i.e. items that express (among other things) the presence of certain syntactic categories, but are not such categories themselves, they are optional—a word need not contain them. They occupy, like most dependents in the syntax of English, the post-head position as the unmarked possibility—they are preponderantly suffixal. Also, as we have observed, syntactic categories may be ‘changed’ in the absence of an affix (as in conversions). Obviously, the suffix in an -able word has to be present to signal the additional component of meaning beyond the category change (and possible incorporation). But it does not itself embody the overall category of the derived word. Such affixes are like such syntactic specifiers as the dab in the German non-finite (38a), where non-finiteness is marked by final position of the verb (compare the verb-second finite main clause in (b)): (25) a. Er sagte, daß er ihn gesehen ha ¨ tte he said that he him seen had (‘He said that he had seen him’) b. Ich hatte den Hut vergessen/Den Hut hatte ich vergessen I had the hat forgotten (‘I had forgotten the hat’) c. Er sagte, er ha ¨ tte ihn gesehen he said he had him seen (‘He said he had seen him’) The specifier is associated with (morphologically finite) syntactically non- finite verb-final subordinates; it is absent in the verb-second (and so finite) subordinate in (25c). -Able is associated with a change in category, but does not itself realize that ‘new’ category. 13.2.5 Absorption, incorporation, and ‘constructions’ All of the forms in (22–24) involve what I’ve been referring to as ‘absorptions’ (elaborated on in §9.2.5). A category is related to a more complex categorial Epilogue: Case, Notionalism, Creativity, and the Lexicon 407 structure which is traditionally said to be ‘derived’ from it; there is ‘addition’ of a superordinate category; compared with the base forms, there is a ‘change’ in category, even if only secondary—as with the concrete count ‘derivative’ of beauty in (23c). In the case of absorptions it is the form expressing the base that is the head of the morphological structure expressing the complex (‘derived’) structure. This reflects overtly the derivation, even (in a sense) in the absence of affixation. But in other instances the absorption structure is associated with a different form from any putative base, as in kill, the causative ‘corresponding to’ die, but not overtly based on it. This is perhaps even plainer with pilgrim, on the assumption that it is verb-based (see §10.1.2). This involves the recognition that lexical structure may be syntactic-categorially complex without this necessarily being signalled overtly, and that this categorial complexity (covert or overt) may have syntactic consequences. Thus, to take a simple example, the {P} that is associated with finiteness formation, whose presence need not be signalled morphologically, provides a free absolutive that hosts subject formation, as in, say, (3a). (3a) also provides a more complex example of the syntactic relevance of (possibly covert) internal categorization, in the form of the structures associated with causatives, lexical or morphological, and their interaction with raising and control, as discussed in Chapters 11 and 12 and above in the present section. I have distinguished these structures and relationships from ‘incorpor- ations’, which may also involve relationships between a simpler and a more complex (‘derived’) categorial structure. In this case the category of the base has subjoined to it a substructure in the ‘derived’ structure, and it doesn’t ‘change’ its category. The discussion of incorporated arguments in the pre- ceding chapters has illustrated these properties. With them what is expressed as the morphological head is also the same as with the base. Thus the passive participle in (3b), for instance is still a verb. Often, incorporations are not overtly signalled morphologically, however, as with contactive formation (9.44) (updated), associated with (9.26b) and (13.28): (9.44)´ Contactive formation {P;N/{abs}{loc}} ⇔ {P;N/{loc,abs}}} | {{abs}} | {N i } 408 Modern Grammars of Case (9.26) a. John supplied the treasure to Bill b. John supplied Bill (with the treasure) The two verb forms are identical. Both incorporations and absorptions are ‘extensions’ of lexical structure, unmediated by the syntax. Absorptions such as we have been looking at provide us with a straight- forward way of accommodating the kind of example that has been used to argue for ‘constructionist’ approaches to relations between lexicon and syntax (as in Goldberg 1995; Goldberg and Jackendoff 2004). Examples such as those in (26) and (27) can be interpreted as absorptions involving a ‘derived’ causative directional verb based on an argument that is in an instrumental relation to it, either a verbal (a) or nominal (b) argument: (26) a. The professor talked us into a stupor b. Bill elbowed his way through the crowd This is indicated schematically in the lexical, so unordered, representation in (27): (27) {P;N} {P;N/{erg}} {N{com},{loc}\{P;N/{erg<abs>}}} | : {P;N/{loc}} : | : {P;N/{abs}{loc}{{erg}}} : : talk/elbow The ‘instrumental’ (‘hybrid’ functor) configuration on the right and the lower {P:N}s joined to the left are subordinate to the highest {P;N}, which is there because of the requirements of a circumstantial, the ‘instrumental’; and the {P;N}s and the ‘instrumental’ are not ordered linearly with respect to each other. The ‘instrumental’ is a comitative ({com}) locative that requires an agentive verb (recall §9.2.3, and particularly (9 .39), updated here as part of (27)). The upper {P;N}s on the left are a causative configuration and the lowest a directional. Thus, in the syntactic structure representing (26a) in (28) the professor satisfies the agentive requirement of the causative, and is hosted by the free absolutives above: Epilogue: Case, Notionalism, Creativity, and the Lexicon 409 (28) {P} | {{abs}} {P;N} : {{abs}} : : {P;N/{erg}} {N{com},{loc}\{P;N/{erg<abs>}}} : | : {{erg}} {P;N/{loc}} : {{abs}} : | : : : {P;N/{abs}{loc{goal}}} : {{(loc),abs}} : : : : : : : : {{abs}} {{loc{{goal}} : : : : : : : : the professor talked us into a stupor And the {{abs}} of the directional (whose spatial source argument is not expressed) is hosted by the (locative) free absolutive of the patient sub- predicator within the causative complex and the free absolutive of the causa- tive ‘action’ predicator itself. The configurationin (28) corresponding to (27) remains unserializable. And the whole complex in (27) i s b ased on (or ‘ deriv ed from ’) the ‘instrumental’ argument (not specified here, as its categories varies, but talk or elbow in the pr esent instances); that is, (27) a ppears on the right-hand side of an absorption relation on the left of which is the lexical representation for talk or elbow,inthecaseof(26). And the whole complex is expressed b y the base form, as in other absorptions. We ha v e a complex c on v ersion. There is no need to appeal to ‘constructions’ with their o wn meaning. This is a relationship between atomic lexical items. Likewise, there is no need to associate the alleged ‘unaccusativity’, or the ‘telicity’, of (29b) versus the ‘unergativity’ (‘atelicity’) of (29a) (cf. German (4.22)) with the distinct contribution of the ‘construction’ in (29b) as such: (29) a. John danced/ran/walked b. John danced/ran/walked to the other side of the room (29b) involves a directional verb ‘derived’ from a simple (‘activity’) agentive intransitive which is, again, in an ‘instrumental’ relation to it. The ‘unaccu- sativity’/‘telicit y’ is associated with the directional verb (recall Keller and Sorace 2003). And again, and as is normal, the base of the absorption, i.e. in this instance the root of the non-directional ‘activity’ verb, is what is ex- pressed overtly. It is unnecessary to attribute to ‘constructions’ properties that belong to the category that projects the ‘construction’. I do not here try to show this on a wider basis. Bo ¨ hm (2001) offers a much fuller discussion of such phenomena, and a rather different interpretation of them, which does 410 Modern Grammars of Case not appeal to an ‘instrumental’ relation (and its apparent problems, related to those discussed by Wunderlich (1997)), but which nevertheless also avoids the ‘constructionist’ conclusion. It is unsurprising if particular kinds of derivational relationship are absent from particular languages, particularly those which are more complex, more marked. Once more, we do not have to associate this with languages having different ‘constructions’, which seems to be an unnecessary complication of the conceptual apparatus of grammar. Languages may have different lexical relationships; there is no need to impose on linguistic variation the compli- cation of ‘constructional’ variation. ‘Constructional’ differences follow from the lexical characterizations of individual lexical items. This is not, of course, to deny that there are multi-word lexical items; many idioms are such. Consider as an example the fell out with verbal sequence in (33), which may be given either an agentive or experiencer interpretation: (30) Colonel Sentence fell out with General Principle The lexical item involved might be represented as in (31), assuming, for illustration, an agentive interpretation: (31) {P;N/{erg}} | : {P;N} : | : {P;N/{abs},{loc{goal}}} {N {com},{loc}\{P:N}} : : : {{loc{goal}} : : : : : : : fall out with From this the syntactic structure in (32) is projected: (32) {P} | {{abs}} {P;N/{erg}} : | : {{erg}} {P;N} : {{abs}} : | : | {{abs}} {P;N/{abs},{loc{goal}}} {N{com},{loc}\{P;N}} : : : {{abs}} : {{loc{goal}} : : : : : : : : : Col. Sentence fell out with Gen. Principle Epilogue: Case, Notionalism, Creativity, and the Lexicon 411 Fall out with is an agentive verb formed on the basis of a directional verb and its goal argument together with a comitative; the agentive argument of the derived verb is linked lexically with the absolutive of the directional. In (34b) the {{erg}} is hosted by the free absolutive of the agentive, and the directional {{abs}} by the free absolutive of the circumstantial predicator; the circum- stantial comitative is hosted by the free absolutive of the agentive predicator. It is only such totally idiosyncratic ‘constructional’ properties, as in (31), that need be entered as part of a lexical item. 13.3 Creativity and notionalism The types of lexical relationship we have been looking at, including the linking mechanism of (10) etc., contribute to linguistic creativity, in enabling metonymic and metaphoric formations. And they thus take us back to a theme of the Prologue, the unacceptability of the per vasive notion of ‘cre- ativity’ voiced by Foley and van Valin as ‘the abilit y of native speakers to produce and understand an (in principle) infinite number of sentences’ (1984: 319) and the inappropriateness of how Chomsky’s (1976) distinction between ‘rule-breaking’ and ‘rule-governed creativity’ is drawn. There are distinctions to be drawn here, to be sure: to do with relative routinization or lexicalization, different dimensions of figurativeness (which is not just a feature of ‘litera- ture’), or what we might distinguish (again following Anderson (1984 c; 1987a)), though scarcely sharply, as ‘suppletive’ versus ‘supplementary’ for- mations. The latter provides alternative means of representing some scene; they are thus usually obviously ‘figurative’ (in a sense, ‘rule-breaking’ or ‘rule- supplanting’), and can lead to ‘idioms’. But the alternative, figurative means of expression can reveal something distinctive about the scene represented; it is to an extent ‘suppletive’, and its content cannot necessarily be identified with any ‘literal equivalent’, even when the metaphor is apparently ‘dead’. The fully suppletive formation provides us with a means of representing a scene for which there is no prior representation (it is ‘rule-creating’, or ‘rule- extending’)—as with the deployment of the localist relations in the represen- tation of abstract as well as concrete ‘spaces’ etc. These localist relations provide ‘literal metaphors’, in another terminology (Lakoff and Johnson 1980, and much subsequent work). As implied, the literal/suppletive/supplementary distinctions are fragile. Man y ‘literal’ expressions are ‘dead’ metaphors (Finally, they understood). And something that might be identified as anidiomatized ‘supplementary’metaphor is merely more recent, and perhaps more transparent, at least vis-a ` -vis its non-literal status, if not in the interpretation of the metaphor (Finally, the 412 Modern Grammars of Case [...]... weight to any and all distributional properties or to all potential members Only the properties of the prototypical use of semantically prototypical members of the category are relevant to identifying the basic distribution of the category Other aspects of distribution correlate 414 Modern Grammars of Case with various sorts of non-prototypicalities of membership or use What we have looked at suggests that... case and alternatives to it, as well as their grounded character, their notional basis Of course, following this evidence has taken us in a number of different directions, some of the more important of which this final chapter has touched on The present book has endeavoured to present something of a history, one based on the consequences of the case grammar hypothesis’, embodying what I called a case. .. inception of transformational grammar the legacy of the tradition of grammars of case This last chapter, however, as well as casting an eye back on this history, has also sought to introduce further consequences of the case grammar’ view, most of them largely still to be developed This is as it should be Every epilogue is also a prologue References ABRAHAM, W (ed.) (1978) Valence, Semantic Case, and... assume, despite the evident groundedness of much of morphosyntax, that syntax must be studied as autonomous Given this, analyses of the syntax of individual languages and of its acquisition will have to appeal to formal devices of such abstractness that the positing of ‘universal grammar’, as the source of such unlearnable ‘abstract principles’, becomes plausible (Anderson 2004a; 2004f) But there is a... LOUNSBURY, F.G (1964) ‘The Structural Analysis of Kinship Vocabulary’, in H.G Lunt (ed.), Proceedings of the Ninth International Congress of Linguists, 107 3–93 The Hague: Mouton LYONS, J (1965) The ScientiWc Study of Language Edinburgh: Inaugural Lecture, University of Edinburgh, no 24 —— (1966) ‘Towards a ‘‘Notional’’ Theory of the ‘‘Parts of Speech’’ ’, Journal of Linguistics 2: 209–36 —— (1967) ‘A Note... to its argument none of the selectional 418 Modern Grammars of Case requirements associated with being subcategorized-for It is required only by the relational character of predication that is articulated by the semantic relations and their mediation between predicator and argument(s) All of this is cognitively salient The rough scenario just outlined remains highly speculative, of course But in this... Wellington University Press ANDERSON, J.M (1968a) ‘Ergative and Nominative in English’, Journal of Linguistics 4: 1–32 —— (1968b) ‘On the Status of ‘‘Lexical Formatives’’ ’, Foundations of Language 4: 308–18 —— (1969) ‘Adjectives, Datives and Ergativisation’, Foundations of Language 5: 301–22 —— (1970) ‘The Case for Cause: A Preliminary Enquiry’, Journal of Linguistics 6: 99 104 —— (1971a) ‘Dependency... far no systematic account in such terms of the prevalence of groundedness, or of the distribution between and within languages of matches and mismatches between syntax and semantics And maintenance of ‘autonomy’ involves both contraction of the traditional bounds of syntax— where it most obviously involves reference to semantics (recall the introduction to Chapter 10) , or to phonology—and expansion (via... 459–576) —— (1981) Review of J.M Anderson (1977), Language 57: 137–40 —— (1985a) ‘The Locus of Case in South Asian languages’, in A.R.K Zide, D Magier, and E Schiller (eds.), Proceedings of the Conference on Participant Roles: South Asia and Adjacent Areas, 211–46 Bloomington: Indiana University Linguistics Club —— (1985b) ‘Mandarin Case Marking: A Localistic Lexicase Analysis’, Journal of Chinese Linguistics... the Role of Thematic Roles in Linguistic Theory’, Linguistics 22: 259–79 CHAFE, W.L (1970) Meaning and the Structure of Language Chicago: University of Chicago Press CHAPIN, P.G (1972) Review of R.P.S Stockwell, P Schachter, and B.H Partee (1968) The Major Syntactic Structures of English (New York: Holt, Rinehart and Winston, 1968), Language 48: 645–67 CHOMSKY, N (1965) Aspects of the Theory of Syntax . a wider basis. Bo ¨ hm (2001) offers a much fuller discussion of such phenomena, and a rather different interpretation of them, which does 410 Modern Grammars of Case not appeal to an ‘instrumental’. the legacy of the tradition of grammars of case. This last chapter, however, as well as casting an eye back on this history, has also sought to introduce further consequences of the case grammar’. its non-literal status, if not in the interpretation of the metaphor (Finally, the 412 Modern Grammars of Case penny dropped). There are often alternative suppletive (or ‘literal’) metaphors that