Desktop version

Home arrow History arrow A Minimalist View on the Syntax–Semantics Relationship: Turning the Mind into a Snowflake

Reversing the direction

The relationship between operations of the syntactic component and its output and the C-I component has been subject to different conceptualizations in the minimalist strand of inquiry into properties of the human language faculty, the diversity itself testifying to the difficulty in investigating components for which, in contrast to components involved in processes of externalization, evidence available for hypothesis testing is mainly indirect and far from unambiguous—a state of affairs particularly uncomfortable for a theory which explicitly rejects common claims about communicative purposes as the driving force for the evolutionary emergence of language, mechanisms of its acquisition and its place among cognitive faculties as an instrument of expressing thoughts:

...there is considerable evidence that (...) fundamental language design ignores order and other external arrangements. In particular, semantic interpretation in core cases depends on hierarchy, not the order found in the externalized forms. If so, then the Basic Property is not exactly as I formulated it before, and as it is formulated in recent literature—papers of mine, too. Rather, the Basic Property is generation of an unbounded array of hierarchically structured expressions mapping to the conceptual-intentional interface, providing a kind of “language of thought”—and quite possibly the only such LOT (...) there is good reason to return to a traditional conception of language as “an instrument of thought,” and to revise Aristotle’s dictum accordingly; language is not sound with meaning but meaning with sound—more generally, with some form of externali- zation, typically sound though other modalities are readily available (...) investigation of the design of language gives good reason to take seriously a traditional conception of language as essentially an instrument of thought. Externalization then would be an ancillary process, its properties a reflex of the largely or completely independent sensorimotor system. (Chomsky 2016b: 13-14)

The primacy of the C-I interface in an explanatory account of the nature and function of the syntactic component has been stressed several times, with the conclusion relevant for the issue of the evolutionary emergence being ‘that language evolved for thought and interpretation: it is fundamentally a system of meaning’ (Berwick and Chomsky 2016: 101), thereby making the nature of the relationship between the syntactic engine and the C-I component the factor relevant for the validity of the Strong Minimalist Thesis (see also e. g. Chomsky (2013c: 36, 39-40), Chomsky (2015b: 5), Chomsky (2015d: 23), Chomsky (2016a: 21-22), Berwick and Chomsky (2016: 74-108)), and for concomitant consequences for the problem of language acquisition, viz. the fact that on this account ‘when you learn your first [language], you’re mainly acquiring the externalization system. (...) The rest you couldn’t be taught, because nobody knows what it is. It’s like being taught how to walk, or something’ (Chomsky 2015a: 74). In sum, ‘the Basic Property of I-language, determined by UG, is a finitely-specified generative procedure, represented in the brain, that yields a discrete infinity of hierarchically structured expressions, each with a determinate interpretation at the CI interface’ (Chomsky 2016a: 22), and the relationship between the former and the latter stands at the center of the search for sources of humaniqueness insofar as it falls within the province of linguistic inquiry. Yet the mutual encounter between narrow syntax and the C-I component as they were seen in the early days of the minimalist enterprise might be rather expected to be full of surprises and troubles with cooperative work: although the modular architecture of the GB period was already gone, together with its hierarchy of levels of representation and a vast array of syntax-specific rules and principles, the repertory of strictly syntactic devices which nevertheless essentially shape objects delivered to the interpretive component and thereby affect the latter remained quite substantive; although the thesis of syntactic autonomy rejects seeing interpretive properties of syntactic objects as being responsible for their behaviour during syntactic computation even in a reduced sense in which uninterpretability of features and the need to eliminate them may be understood as a semantics-oriented property, this does not make outputs of the derivational process less involved in the workings of the C-I component. The reliance on purely syntactic features, together with the claim— known as the Strong Minimalist Thesis—that ‘language is an optimal solution to legibility conditions’ (Chomsky 2000a: 96) and apparent richness of interpretive resources at the disposal of the C-I component led to the ‘matching complexity issue’: ‘what the syntax qua computational system of language produces has to match what is independently there, in the hypothesized pre-linguistic conceptual- intentional (C-I) systems’ (Hinzen 2011a: 423). The evolutionary emergence of narrow syntax, involving on the standard minimalist picture much more than a mechanism to build hierarchical objects—even if its sole structure-building operation is merge, its non-local mate, Agree, operates on a substantial collection of properties, which stand in need of being valued and eliminated before reaching the interface with the C-I system—creates already problems of its own; its matching an independently present interpretive component would verge on a miracle:

Grammar as founded upon Merge and the need to eliminate such ‘features’ has no inherent motivation or foundational significance: it solves a problem it has created for itself, through the existence of purely formal features that need to be eliminated. On this picture the lexicon is inherently motivated, insofar as it interfaces with the semantic component directly. Grammatical representations, on the other hand, have to be ‘cleansed’ before such an interface can be accessed (the problem of grammar, as it were, is grammar itself ). This architecture makes sense on the assumption of an autonomous syntax and the independence of thought (C-I systems) from grammar. But it becomes unclear, on this view, why there should be a lexicon-grammar distinction in the first place—or why we ever leave the realm of the interpretable, and enter the realm of the grammatical. (Hinzen and Sheehan 2013: 87)

The alternative seems to be to reverse the direction of explanation and to let semantic properties be dependent on strictly syntactic properties in a way which would make the C-I component being shaped by narrow syntax instead of the latter satisfying the needs of the former, a research program articulated and defended at length in particular in Hinzen (2011a,b, 2012a,b,c,d, 2013), Hinzen and Sheehan (2013); Sheehan and Hinzen (2011), Hinzen, Sheehan, and Reich- ard (2014): ‘these two options could not be more different: in the former, Chomskyan option, syntax answers semantics (or expressive needs); in the latter, it is the other way around’ (Hinzen 2012b: 131). Drawing the dichotomy in sharply delineated terms helps making differences clear, but certain caveats may be required in order to take into account the most current picture of the properties of narrow syntax and to avoid falling into a trap of attributing to narrow syntax effects which should be rather ascribed to a shared effort of narrow syntax, semantics as it arises at the syntax-the C-I component joint, and further cognitive components. Explaining the appearance of the systematic availability of notions of events, facts and propositions as their being founded on the availability of phase-based syntactic objects and a hierarchy of their complexity runs the risk of confusing correlation and explanation as much as it was argued to happen in the opposite direction with Chomskyan explanation of the duality of semantics in terms of interface requirements:

[Chomsky] does make a relatively specific proposal for a way in which configurations and operations in narrow syntax correlate with aspects of semantic interpretation: external Merge with argument structure, internal Merge with discourse semantic and information- structural aspects. Yet, this account assumes, with many semanticists, that the semantic duality in question exists on the non-linguistic side of the interface, independently of which syntactic configurations the grammar delivers there, so as to constrain the operations of narrow syntax. Hence the duality (or semantics as such) is not explained. Generally, correlation as such is not explanation, and the tighter the correlation, the more the question arises whether the duality that we unquestionably find on the grammatical side is simply re-described in non-grammatical (i.e. semantic) terms when it is said to also exist on the other (non-linguistic) side of the ‘interface’. (Hinzen 2012d: 638)

It is certainly true that the assumption of ‘sufficient diversity’ of expressions delivered by narrow syntax to the C-I component relies too much on tacit assumptions about quite specific properties of the latter, as the requirement of interpretive articulation in terms of information structure; but, turning the tables, it does not seem to offer much better prospects for explanation to endow grammatical structures as such with interpretive import of this sort—as Boeckx (2015b) notes with regard to the revamping of modi significandi proposed in Hinzen and Sheehan (2013),

...eventually the causal chain of explanation must be broken. If it’s not a feature that defines a grammatical category, but a mode of signifying, and if we say that we have the mode of signifying we do because of the grammatical categories we have, what provides the foundation to what? Chickens and eggs indeed. (Boeckx 2015b: 168)

Difficulties in reducing complex interpretive properties to properties of syntactic structures become only more severe in the demanding framework of syntactic theory as it emerges in Chomsky (2013c, 2015b), in which not only is the characterization of properties of syntactic objects in semantic terms avoided (the most crucial cases being the notions of phases and of unvalued features, in both cases understood purely formally; see section 3.4.2), but also such mechanisms as projection and properties like categorial labels are eschewed, used only for convenience to indicate kinds of syntactic objects manipulated during derivation and interpretation without recourse to such concepts. On the other hand, the most recent developments in syntactic theory actually lessen the difficulty of the match between syntax and the C-I component. First, the impoverishment of the syntactic apparatus reduces theoretical complexity on the syntactic side. Second, clarifications of the Strong Minimalist Thesis shed more light on its conceptual content:

If FL is perfect, then UG should reduce to the simplest possible computational operation satisfying the external conditions, along with principles of minimal computation (MC) that are language-independent. The Strong Minimalist Thesis (SMT) proposes that FL is perfect in this sense. SMT is not precisely formulated. MC can be interpreted in various ways, though some of its properties are uncontroversial, and reliance on these carries us a long way, as work stimulated by MP has shown. There is a plausible suggestion as to what the simplest computational operation is: Merge, as defined within MP. SMT accords with the guiding principle of the natural sciences, and there is reason to expect something like this to be correct on evolutionary grounds. (Chomsky 2015e: ix)

The notion of ‘external conditions’ in the case of the C-I component, the primary one on current assumptions, is not only less precise than the concept of minimal computational principles given the speculative nature of hypotheses about properties of the C-I system as such; it also seems that it is not the notion that should be taken to be relevant for an investigation of the workings of narrow syntax and the C-I component as a default source of explanation:

In the best case, phenomena would be explained by interaction of the simplest computational operation—Merge, with its two logically possible subcases, Internal Merge IM (automatically yielding “the copy theory of movement”) and External Merge EM—interacting with general principles of minimal computation MC. The Strong Minimalist Thesis SMT articulates this goal. (Chomsky 2015b)

Instead of choosing between a picture in which the C-I component ‘shapes’ narrow syntax in that it imposes legibility conditions on the output of the latter, conditions which would ultimately reflect its internal richness and complexity, and a view according to which narrow syntax ‘shapes’ the C-I component in that it is endowed with intrinsic meaning and the latter emerges from this encounter essentially changed, blindly following the path that the former carved, to appropriate the catchy phrase of Uriagereka (1999: 275), one may consider abandoning the dichotomy and, while accepting ‘semantic blindness’ of syntax, as Hinzen (2006) calls freedom of narrow syntax from semantic influence, and the syntax-semantics direction of molding, one could walk a middle way of trying to discern properties and operations of the C-I component which are indeed directly shaped by narrow syntax insofar as it provides hierarchically structured objects for interpretation, and those interpretive properties and devices that utilize whatever happens as an effect of the immediate contact of syntax and semantics and build upon these results, combining them possibly with outputs of other C-I related components. The search for minimal semantics in this sense starts with properties of syntactic structures, taking into account a radically reduced collection of theoretical concepts available in the most current formulations of the minimalist program, and strives to model those interpretive properties and procedures which neither ignore syntactic properties nor introduce mechanisms and concepts ex machina, both narrow syntax and the interpretive (subcomponent being expected on this account to have properties ‘explained by interaction of the simplest computational operation (...) interacting with general principles of minimal computation MC’. Sharpening of the Strong Minimalist Thesis to make the principle of minimal computation the relevant principle guiding explanatory accounts of syntax easily admits its extension to the syntax-semantics interaction, thereby restricting and elucidating the notion of ‘syntax giving shape to semantics’. It allows to avoid equating the human faculty of language and capability of thinking in general without giving up a project of finding sources of humaniqueness in the former. Following a metaphor used to explicate properties of narrow syntax expected on minimalist grounds, the question ‘To what extent is it like a snowflake and to what extent is it like a spine?’ (Chomsky 2004b: 156), asked about the C-I component, receives as a general answer a statement that the C-I side is like a snowflake to the extent that narrow syntax influences it; and properties going beyond those following directly from properties of syntactic objects, while not excluded in principle, should be admitted only if it is unavoidable, preferably being attributed to later stages of processing in the cognitive system. While ‘C-I must have some range of resources that can exploit the properties of generated expressions, along with whatever is involved in use of language to reason, refer, seek to communicate perspicuously, and other mental acts’ (Chomsky 2007: 15-16), resources of which the interpretive component avails itself for the purpose of interpreting syntactic objects as they are delivered from narrow syntax might be hypothesized to be essentially more restricted than ‘whatever is involved in use of language to reason, refer, seek to communicate perspicuously, and other mental acts.’ ‘Use of language to reason, refer, seek to communicate perspicuously, and other mental acts’ does not fall within the province of the science of language tout court; interpretive properties of syntactic objects and processes to which they are subject at the C-I side of the derivational process do. All this neither eliminates questions about properties specific to the C-I component which, while not being founded upon properties of syntactic objects, are nevertheless required for interpretation to proceed; nor does it make the inquiry into relationships between interpretive procedures so understood and cognitive consequences of having narrow syntax illegitimate—rather, it helps distinguish different phenomena to be modeled. Simple correlations between objects formed by applications of merge and sophisticated concepts and cognitive tools which seem unique to the human species are hardly expected to be found—but this only means that the picture might become more detailed and nuanced. Instead of finding complex abstract concepts directly encoded in syntactic properties and structures, we may find mechanisms which let such uniquely human cognitive apparatus grow.

 
Source
< Prev   CONTENTS   Source   Next >

Related topics