How Can Semantics Avoid the Troubles with the Analytic/Synthetic Distinction?

At least since Quine (From a logical point of view. Harvard University 1 Press, Cambridge, MA, 1953) it has been suspected that a semantic theory that rests 2 on defining features, or on what are taken to be “analytic” properties bearing on the 3 content of lexical items, rests on a fault line. Simply put, there is no criterion for 4 determining which features or properties are to be analytic and which ones are to 5 be synthetic or contingent on experience. Deep down, our concern is what cogni6 tive science and its several competing semantic theories have to offer in terms of 7 solution. We analyze a few cases, which run into trouble by appealing to analyticity, 8 and propose our own solution to this problem: a version of atomism cum inferences, 9 which we think it is the only way out of the dead-end of analyticity. We start off 10 by discussing several guiding assumptions regarding cognitive architecture and on 11 what we take to be methodological imperatives for doing semantics within cognitive 12 science—that is a semantics that is concerned with accounting for mental states. We 13 then discuss theoretical perspectives on lexical causatives and the so-called “coer14 cion” phenomenon or, in our preferred terminology, indeterminacy. And we advance, 15 even if briefly, a proposal for the representation and processing of conceptual content 16 that does away with the analytic/synthetic distinction. We argue that the only account 17 of mental content that does away with the analytic/synthetic distinction is atomism. 18 The version of atomism that we sketch accounts for the purported effects of analyt19 icity with a system of inferences that are in essence synthetic and, thus, not content 20

(1) Mary began a book. 58 Imagine now that the issue at hand is how a sentence such as (1) might be inter-59 preted. The proposal quoted in (2) is apropos the sorts of psychological events carried 60 out during the comprehension process of (1). The semantic issues underlying this 61 proposal will be dealt with a little later, but we start off with the commitments of this 62 proposal vis-à-vis cognitive architecture. 63 (2) "(a) When encountering the noun book, comprehenders access the word's lexical entry and attempt to integrate various stored senses of this word into the evolving semantic representation of the sentence.
(b) The mismatch between the verb's selectional restrictions and the stored senses of the noun triggers a coercion process.
(c) Comprehenders use salient properties associated with the complement noun and other relevant discourse elements (including but not necessarily limited to the agent phrase) to infer aplausible action that could be performed on the noun. The postulation that the semantic representation of token lexical items varies along n dimensions (multiple features or senses or salient properties) is incompatible with the view that semantic representations are symbolic and that processes are computational. Symbolic expressions and operations over them require that constituent symbols hold across contexts and expressions-thus require a "classical" notion of compositionality. If token items were to vary along numerous dimensions, in the extreme case, contradictions (P &~P) could be taken as valid, depending on which properties were to contribute content to the constituents of the two P's.

490394_1_En_5_Chapter
TYPESET DISK LE CP Disp.:23/10/2020 Pages: 21 Layout: T1-Standard process of understanding a sentence; they are the same we ought to use in semantic 122 analysis. As such, we assume that in order to account for the nature of these cognitive 123 processes-that is, in order to account for the nature of those thoughts-it is crucial 124 we not only understand the nature of the elementary parts, but also how they combine 125 to yield the meaning that the thought carries. 126 Moreover, we think that to entertain a thought is to entertain something like 127 a proposition whose basic elements are concepts. We take a proposition to be a tions between conceptual constituents (see Kintsch, 1974;and McKoon & Ratcliff, 135 1992, for early propositional theories). This view also applies to the process of 136 language comprehension: understanding a sentence requires recovering the meanings 137 of words/morphemes in the context of the proposition that the sentence expresses.

138
Propositions are thus the mental objects whose referents are states and events in the 139 world (and ideas about events and states in the imaginary world, if you will). In 140 order for propositions to refer, or in order for propositions to stand for the events and 141 states whose contents they represent, they have to compose, and in order for them to 142 compose they require a syntax. between semantics and pragmatics (for further discussion, see de Almeida, 2018).

201
As we will see, a key issue-in line with that we see in proposal (2)-is the idea of 202 "coercion". We turn to these matters now. be clear, before we advance discussion, that our concern is not with truly analytic 220 statements such as those in which a conjunction entails its parts. These are run over 221 form-something like P&Q → P. is with other, often subtler, forms of analyticity, common to lexical-semantic theories 226 as well as theories of composition relying on certain types of semantic operations 227 such as "coercion". And, more broadly, our main concern is with the shaky ground 228 upon which all of semantics that appeal to analytic features stands.

229
There are, we think, roughly three ways to conceive how a concept might enter 230 into-i.e., contributes content to-a proposition. (i) The first is by contributing its full 231 content, whatever that may be. If one believes concepts to be composed of particular 232 sets of features, then the content that a given concept contributes to a proposition must 233 necessarily be that particular set of features-nothing more, nothing less. (ii) Another 234 way in which a concept might contribute content to a proposition is by contributing 235 some, but not necessarily all, of its features. If one believes a concept to be made 236 up by a set of features, then, the kinds of features that a concept might contribute to 237 a particular proposition is relative to the particular context of the proposition-that 238 is, it is sensitive to other constituent concepts, perhaps to the wider discourse, and 239 perhaps to the syntax of the expression. And (iii) the third way in which a concept 240 3 An anonymous reviewer was right at pointing out, among other problems, that the analytic/synthetic issue that we are trying to "reawaken" is "not new". This, of course, is not an argument against our view. If anything, this is an embarrassment for semantic theories. We believe that the two case studies we discuss below, though limited in scope, are representative of a widespread practice in semantics. It should be noted that the kind of a/s issue we are raising is about mental representation, not linguistic analysis. a concept has no features. In the present section, we will discuss (i) and (ii); the case 243 for (iii) will be further advanced in Sect. 3. 244 We cannot possibly be exegetic in our evaluation of semantic theories that are 245 committed to analyticity (see, e.g., Engelberg, 2011a, for review). Our goals here 246 are to illustrate the state of the art and thus motivate our proposal for moving away 247 from analyticity-namely, to make the case for our brand of atomism. And we will and their carrier sentences. 285 We assume that semantic templates such as (3b) are intended to represent the 286 propositional content of (3a) specifying its form and key elements of meaning. We are assuming throughout that these theories all postulate that template structures are representations of psychological objects, as in Jackendoff (1983), similar to representations in a language of thought, though this is not always explicit in the works we cite. 5 Although most of our discussion focuses on a theory such as Levin and Rappaport Hovav's (2005), we assume that the main points we make apply to all theories we mentioned. 6 An anonymous reviewer pointed out that, "Most people don't assume that in order for there to be synonymy (and thus, analytic truths), the expressions in question need to be psychologically perfectly equivalent. For instance, it is standardly accepted that a correct analysis can be highly nonobvious." We fail to understand what "most people" assume, for we do take synonymous sentences in natural language to be expressions of "perfectly equivalent" mental states (viz., propositions). As a further example of this state of affairs, consider the distinction between 348 so-called "externally caused" and "internally caused" change of state verbs such as 349 those in (5a) and (5b) respectively. Before we further explore this issue, in contrast to atomism in Sect. 3, we would 375 like to address rather briefly a second semantic phenomenon-coercion-one for 376 which appeals to analyticity are also quite evident. 378 The term "coercion" (or type-coercion, or type-shifting) is identified with partic-379 ular hypotheses on how sentences such as (1) are interpreted-among which is the 380 proposal presented in (2). We refer to these sentences as "indeterminate" because

397
There is perhaps some confusion here between meaning, sense, and use-damage 398 that unfortunately Wittgenstein cannot come back to repair. If we tell you that it is hot 399 today, in Montreal, when actually it is −20°C, we are most likely being sarcastic.

400
It does not entail, now, that the concept HOT includes COLD, among its senses. 401 We are certainly using the word hot to convey something else entirely, to provoke 402 you or, as Davidson (1978)  relations. And, throughout, we tailor our discussion of atomism and inferences to the 427 analysis of the two phenomena we discussed in Sect. 2. 428 We have mentioned that we are committed to symbolic representations and to 429 computational processes. Patently, we take symbols that stand for content to be things (and events) they refer to. We do not establish a lower limit on the content 435 that the simplex symbols convey-or more properly on the very content that they 436 individuate-but we suggest that they are properties, predicates, and "particulars", 437 as Russell (1913) once put it. We assume that, for the most part, atoms are expressed 438 by the simplex bound and free morphemes of natural language. And since we take 439 concepts to be the very symbols of (again, Russell) our "experience", we assume that 440 they enter into different cognitive processes via computations.

441
So much for linking our view of conceptual representation and processes to the 442 architecture we presented in Sect. 1. As for the nature of conceptual representation, if 443 concepts are "atoms", they are simply individuated by the kinds of things they refer.

444
One quick note should suffice to address the problem of reference here: while we 445 take concepts to be pointers to objects (in a very broad sense, including properties 446 like patches of color) and events, they are also representations of things for which 447 there is no referent (or, again, as Russell put it, in the "past, present, or not in time at 448 all", p. 5).

449
Two further observations are in order. The first is that it is likely that the things In our system, the relation is inferential, not one of dependency. 9 If so, most likely 463 the kinds of "particulars" that the conceptual system locks into may be the very entry 464 points to the sets of inferences one runs in conceptual processing. This may become 465 clearer with an example.

466
Take (6) to be the referential relation that obtains between the word (or the object) 467 dog and its concept.
The locking mechanism that affords DOG out of the word or object is a mechanism 470 that in principle is tokened by whole objects, assuming that the visual attentional 471 mechanism locks into full objects (see Fodor & Pylyshyn, 2015;Jackendoff, 2002  their actual content cannot be accounted for by semantic analysis. 518 We also acknowledge that even those with whom we share the main tenets of 519 atomism have argued against adopting MPs for they are too unconstrained and thus 520 cannot be used as an account of semantic inferences (Fodor, 1998 530 We turn now to the other phenomenon, that of the comprehension of indeterminate 531 sentences such as (1). To ease discussion and comparison with (2)-we will cast our 532 proposal rather informally as in (8).

533
(8) (a) Every incoming token lexical item (i) maps onto its corresponding concept (book BOOK), (ii) contributes its syntactic information (book __N ), and (iii) contributes logical (viz., a logical form; (b) The evolving syntactic parsing for a sentence such as (1) tags all its lexical constituents and its linguistically motivated gaps-viz., the gaps for syntacticpositionsthatmaybe optionallyfilled-inlexically.
Asfor (1) (d) The meaning of a sentence is obtained by combining the token concepts-the translations of morphemesy into the evolving (=BOOK) (BEGIN (x, y)) (or, alternatively, (BEGIN (x, y, w))); that is the shallow, (e) Many processes of enrichment ensue; among them are the processes of filling the gaps identified during syntactic structuring with the concepts that were part of the postulates triggered by (i) the utterance context, and (ii) the co-text.
x BOOK(x)

VP [V 0 began
to the evolving syntactic tree information to an evolving semantic composition [ x, BOOK(x)]).
form, such as x(=MAN), w logical unenriched interpretation of (1). 534 We can only make brief observations about (8)-but we trust that the contrast 535 with (2) is quite clear. First, notice that the meaning of book is not a sense; and, 536 according to our proposal, there are no senses stored with the meanings of words. 537 We do not deny that there are uses, but uses are obtained pragmatically (they are 538 synthetic; see below), within the inferences that run after conceptual tokening (as 539 in 8a) and conceptual composition. Also, as suggested in (8b)  The advantage of a proposal such as the one sketched in (8), in summary, is that 545 it does away with analyticity. For any of the proposals appealing to analytic proper-546 ties, the burden is to determine the criterion for separating analytic from synthetic 547 properties. We do not appeal to such properties because to us concepts are atomic, 548 but we see a role for such properties in the inferences that ensue upon conceptual 549 tokening and semantic composition. we can dispense with a semantics that attempts to legislate on experience and world 561 knowledge.

562
Second, we assume that many of the inferences that run as a consequence of 563 a concept being triggered are common to many inhabitants of the same commu-564 nity, those sharing similar kinds of experiences. We cannot be precise on this idea 565 because it points to something whose variables are virtually infinite. Crucial to our 566 approach, in fact, is the idea that these commonalities cannot be legislated on. We 567 also suggest that many, perhaps most effects found in the literature-from priming to 568 prototypicality-are manifestations of these inferences; they are effects of the causal 569 connectedness established between concepts as a function of use and experience. And 570 we even acknowledge that it may be difficult to dissociate-empirically-between 571 inferences computed upon tokening concepts and effects of "activation" of prop-572 erties. However, we have presented some clear signs from the literature that point 573 against decomposition. 574 We do hold that there is a crucial distinction, upon which a theoretical advantage 575 stands: by not taking properties to be analytic, there is no commitment to building 576 a semantic theory whose foundations are faulty. The crucial distinction between 577 atomism and molecularism is that the former, but not the latter does not require 578 semantic analysis based on features or synonymy and, because of that, there is no 579 analysis of content other than assuming that concepts (and their lexical labels) are 580 largely referential, symbols that point to things, events, ideas, and so forth. Reference If semantics appeals to features, without an analytic/synthetic distinction, it turns 584 to holism, which is the antithesis of semantics-at least of a semantics committed to 585 compositionality and productivity. If semantics appeals to properties of the world to 586 fix properties of mental representations, it may fall into the intentional fallacy trap.

587
The way semantics can avoid all this trouble is to turn to atomism cum inferences. Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.