LINGUIST List 30.2418

Tue Jun 11 2019

Review: Pragmatics; Semantics: Sawada (2018)

Editor for this issue: Jeremy Coburn <jecoburnlinguistlist.org>



Date: 25-Nov-2018
From: Brady Clark <bzacknorthwestern.edu>
Subject: Pragmatic Aspects of Scalar Modifiers
E-mail this message to a friend

Discuss this message


Book announced at https://linguistlist.org/issues/29/29-837.html

AUTHOR: Osamu Sawada
TITLE: Pragmatic Aspects of Scalar Modifiers
PUBLISHER: Oxford University Press
YEAR: 2018

REVIEWER: Brady Clark, Northwestern University

INTRODUCTION

Sentences can be used to express multiple propositions. A speaker who utters ''Sam regrets drinking yesterday” seems committed to the propositions that Sam drank yesterday and that Sam regrets it. The second proposition is likely more conspicuous than the first. A great deal of recent research has been devoted to this multidimensionality.

Central to this research is the notion of (not-)at-issueness. At-issue propositions are often described as the ''main point'' of the utterance and not-at-issue propositions as ''backgrounded''. (Not-)at-issueness is dependent on several factors, including convention and discourse. Ideally, characterizations of (not-)at-issueness determine diagnostics for (not-)at-issueness and the assumptions that underlie them. But research on (not-)at-issueness ''did not proceed in this orderly fashion and formal characterizations of at-issueness that go beyond calling at-issue content the 'main point' and not-at-issue content 'backgrounded' … have only recently been developed'' (Tonhauser et al. 2018: 528).

Sawada (2018), the book reviewed here, investigates scalarity, focusing on Japanese scalar modifiers. Roughly, the content of a scalar expression maps its argument(s) onto a set of points ordered along some dimension (e.g., height). For instance, the adjective ''tall'' expresses a function from objects to degrees, the degree to which the object is tall (at a certain time, location, …). Many kinds of expressions express scalarity. Sawada (pg. 1) claims some scalar expressions are ''dual-use'', ''where a degree morphology/scalar concept used for expressing at-issue scalar meaning … can also be used for expressing a not-at-issue scalar meaning.'' He examines multiple scalar expressions in a ''multidimensional composition system'' (pg. 8).

Sawada's book is a significant contribution to the literature. He presents a comprehensive description of a class of scalar expressions. This book will be valuable for researchers interested in the expression of scalarity in Japanese. Linguists interested in not-at-issue content will benefit from Sawada's investigation. But there are problems with Sawada's descriptive claims and theoretical proposals, some particular to Sawada's book, some endemic to the (not-)at-issueness literature. These problems illustrate challenges associated with analyzing (not-)at-issue content. To keep what follows to a reasonable length, I focus the main body of the review on how Sawada diagnoses not-at-issue content.

SUMMARY

Chapter 1 is the introduction. Chapter 2: Sawada’s focus is conventional implicature (CI) scalar content. He adopts Potts’s (2005) characterization of CIs. A CI is not-at-issue content that comments on at-issue content. Unlike conversational implicatures, CIs are semantically encoded. For example, appositives like ''the dancer'' in ''Chris, the dancer, shouted” encode CIs (Potts 2005). Sawada's primary claim regarding CI scalar content is that it is ''cross-dimensional'' (pg. 35). Japanese scalar expressions like ”totemo'' ('very') are “dual-use” expressions that encode at-issue and not-at-issue content which share a single semantic structure.

Chapter 3: Sawada presents his characterization of CIs, building on Potts (2005) and McCready (2010). Different semantic types distinguish at-issue content and CIs. A sentence encoding a CI gives rise to at least two propositions (Potts 2005); e.g., a speaker who utters ''Sam, who lost millions, wept'' (containing the appositive ''who lost millions”) in response to ''How’s Sam?” is committed to the at-issue proposition that Sam wept and the CI that Sam lost millions.

Chapters 4-7 investigate different scalar modifiers. Sawada proposes two types of scalars: higher-level ones, which utilize pragmatic scales, and lower-level ones, which recycle (pg. 8) the scale of an at-issue predicate. Chapters 4 and 5 concern higher-level scalar modifiers; Chapters 6 and 7, lower-level. Chapter 4 explores the pronoun ''nani-yori-mo'' 'more than anything else’, which expresses two kinds of comparison with indeterminateness. On the “individual” reading, content like tennis is compared with a contextually determined set of alternatives. On the ''noteworthy” reading, a proposition is compared with alternative propositions in terms of noteworthiness. Sawada argues that the noteworthy implication is a CI and the individual implication at-issue, while positing scale structure sharing.

Chapter 5 investigates polarity items ''chotto'' and ''sukoshi'', both meaning, roughly, ‘a little'. Sawada argues they have the same at-issue content but contribute different CIs: ''sukoshi”’s CI content is that the speaker's measurement is precise, while ''chotto'''s is that the measurement is imprecise. Sawada claims ''chotto'' but not ''sukoshi'' is used as a speech act minimizer whose CI is, roughly, ‘the degree of imposition of the assertion on the hearer is slightly greater than the minimum.’

Chapter 6 focuses on the intensifiers ''totemo'' and ''motto'' (roughly, 'even more', i.a.). For CIs, both expressions use the scale encoded by a gradable predicate. ''Totemo'' combines with a negative gradable expression to conventionally implicate a high degree of unlikelihood. ''Motto'' combines with a gradable predicate to conventionally implicate that the expected degree (of, e.g., deliciousness) is lower than the actual.

Chapter 7 explores two adverbs, ''yoppodo'' and ''kaette''. “Yoppodo'' communicates an intensified meaning; e.g., when ''yoppodo'' combines with the adjective ''oishii'' 'delicious', 'very delicious' is conveyed at the at-issue level, with a counter-expectational CI that the target’s degree is above the judge's expectation. In contrast, ''kaette'' expresses an at-issue identity function and the CI that it is normal that the property P (e.g., 'dangerous') doesn’t hold.

Chapter 8 examines subject- and speaker-oriented implications triggered by embedded scalar modifiers, focusing on belief-predicate complements. Sawada argues subject-oriented readings are available for many embedded scalars. Scalars vary in what type of environment supports subject-oriented vs. speaker-oriented readings; e.g., ”yoppodo'' requires a main clause evidential modal for the speaker-oriented interpretation. Sawada proposes that, for subject-oriented readings, CIs triggered by scalars undergo a semantic shift (not-at-issue to at-issue content).

The historical development of scalar modifiers is Chapter 9’s focus. Sawada proposes that most scalars tend to be the result of shift from propositional to expressive meaning. He considers the interaction of syntactic and semantic change, arguing that semantic change is sometimes not accompanied by syntactic.

Chapter 10 summarizes the book and compares Sawada's proposals to others (e.g., Bach 1999).

CRITICAL EVALUATION

The notion of (not-)at-issueness is central to much recent work, primarily because of the impact of Potts (2005). Potts, however, never provides an explicit characterization of at-issueness (Amaral et al. 2007: 729). This is also true of much work subsequent to Potts (2005), including Sawada’s book. Some exceptions include AnderBois et al. (2015), Murray (2014), and Simons et al. (2010).

Beyond an explicit characterization, work on (not-)at-issueness should also involve consistent use of diagnostics for varieties of (not-)at-issue content. Sawada inconsistently uses several diagnostics to support his claim that some scalar modifiers are associated with not-at-issue CIs. This inconsistency is instructive, demonstrating difficulties associated with diagnosing (not-)at-issue content.

Amaral et al. (2007: 711-712) observe that ''the main division of content [in Potts 2005] is determined by the notion of 'deniability''', where deniability is susceptibility to direct rejection by ''(the relevant language's equivalent of) 'That is (not) true''' (Matthewson et al. 2007: 220). It is often assumed that direct rejectability is a sufficient property of at-issueness: If a proposition p can be directly rejected, then p is at-issue.

(1) illustrates the application of direct rejection to an utterance that contains an appositive (''a dancer'' in A’s utterance). The direct negative reply B, challenging the main clause content of A (i.e., that Sam took the bus), is felicitous. B', which targets appositive content, isn’t. Assuming direct rejection diagnoses at-issueness, the infelicity of B’ suggests that appositive content is not at-issue.

(1) A: Sam, a dancer, took the bus.
B: False! He didn't take the bus.
B': #False! He isn't a dancer.

Sawada presents direct rejectability as a hallmark of at-issueness (pg. 3). He shows that only certain uses of scalar modifiers can be targeted by direct rejection. We should be cautious, though, about using direct rejection as an at-issueness diagnostic (see Snider 2017). But whatever the viability of the direct rejection as an at-issueness diagnostic, it doesn’t distinguish between presuppositions and CIs. Both resist direct rejection.

Sawada also discusses the ''Wait a minute'' test (Shanon 1976, von Fintel 2004, Potts 2008), a diagnostic that is sometimes used to show that, although a certain speaker commitment cannot be directly rejected, it can be challenged indirectly with expressions like ''Wait a minute”:

(2) A: Sam, a dancer, took the bus.
B: Wait a minute. He isn’t a dancer.

Sawada assumes that this test targets presuppositions. Potts (2008) and Syrett and Koev (2015) provide evidence against this assumption, showing, for example, that “Wait a minute'' can also challenge appositive and main clause content. The ability to pass the ''Wait a minute'' test is a necessary, not sufficient, condition for presuppositionhood (Potts 2008).

Sawada discusses (pgs. 36-38) two properties he claims distinguish CIs from presuppositions: projectivity and backgroundedness. For Sawada, projectivity concerns the interpretation of an implication embedded under a belief- or speech-predicate. The interpretation of the appositive ''the linguist'' (that Kim is a linguist) in the following is preferentially speaker-oriented:

(3) Sam believes that Kim, a linguist, loves to dance.

On Sawada's definition, content is “projective” when embedded under a belief-predicate if the embedded content is speaker-oriented. Sawada's conception of projectivity is different from others in the literature. Tonhauser et al. (2013: 66-67) characterize an implication as projective if the implication tends to be understood as associated with the speaker ''even when the trigger is embedded under operators [like negation] that usually block the implications of material in their scope.'' These environments include negation, polar questions, epistemic possibility modals, and the antecedent of a conditional. I think Sawada’s “projectivity” is related to what Tonhauser et al. (2013: 93) call the “obligatory local effect'':

OBLIGATORY LOCAL EFFECT: A projective content m with trigger t has obligatory local effect if and only if, when t is syntactically embedded in the complement of a belief-predicate B, m necessarily is part of the content that is targeted by, and within the scope of, B.

Sawada claims (pg. 36) presuppositions cannot ''usually'' project out of ''the complement of an attitude predicate/a verb of saying, CIs can'' (Potts 2005; see Heim 1992 on projection in these contexts).

Sawada presents evidence that the not-at-issue content of some embedded Japanese scalars are projective, but typically only under certain conditions. For example, “motto'' allows for both speaker-oriented and non-speaker-oriented readings under belief-predicates, if there is a deontic modal in the main clause (pgs. 182-183). In this way, the implication triggered by “motto” behaves quite differently than appositive content in English, which is typically speaker-oriented.

The presuppositions triggered by some classical presupposition triggers are known to project out of belief-predicate complements (Heim 1992). Tonhauser et al. (2013) provide evidence that the implications associated with certain expressions that have been classified as presupposition triggers --- for example, the possessive relation expressed by possessives and the salience implication associated with the additive ''too'' --- are not necessarily part of the attitude-holder's belief state. Instead, they can be associated with the speaker. Projectivity (in Sawada's sense) does not neatly distinguish CIs from presuppositions.

The second property that Sawada uses to distinguish CIs from presuppositions is backgroundedness. Potts (2005: 33-34) observes that, e.g., appositives, seem unable to express information already in the common ground. Potts (2005) proposes an ''antibackgrounding'' requirement that holds for at-issue content and CIs but not presuppositions. Consider the continuations in (4i) and (4ii) (adapted from Potts 2005: 34). (4i), which contains an appositive (''a cancer survivor''), is infelicitous. In contrast, (4ii), in which the proposition that Lance Armstrong is a cancer survivor is a not-at-issue implication triggered by ''know'', is felicitous.

(4) Lance Armstrong survived cancer.
(i) #When reporters interview Lance, a cancer survivor, he often talks about the disease.
(ii) And most riders know that Lance Armstrong is a cancer survivor.

Sawada's claim is that the not-at-issue scalar content triggered by scalars is a CI. He does not always provide clear evidence for this. He asserts, e.g., that polarity items ''chotto'' and ''sukoshi'' trigger an imprecise and precise CI, respectively (pgs. 84). Sawada's evidence (pg. 92-93) for the not-at-issue status of these implications comes from two sources. First, the implications evade direct rejection. Second, the implications “do not logically interact with logical operators”: the imprecise/precise implications for these two minimizers tend to be understood as associated with the speaker even when these minimizers are embedded within a question or the antecedent of a conditional.

Having provided evidence that the implications associated with ''chotto'' and ''sukoshi'' are not-at-issue, Sawada must establish that the implications are CIs. The evidence he provides for this claim comes from projectivity: the not-at-issue content triggered by these two minimizers can project past a belief-predicate. But, as discussed above, projection out of a belief-predicate complement is not a property that cleanly distinguishes presuppositions from CIs. As Sawada notes (pg. 93), the implications triggered by ''chotto'' and ''sakoshi'' can be associated with the attitude-holder and the speaker. Sawada also asserts that the not-at-issue implications associated with ''sakoshi'' and ''chotto'' are ''new information''. But he provides no linguistic evidence for this.

A similar lack of thoroughness characterizes Sawada's discussion of “totemo”. Sawada provides evidence (pg. 110-111) that the implication triggered by ''totemo'' is not-at-issue: the implication cannot be targeted by direct rejection and “cannot be under the scope of logical operators” (pg. 110). But he provides no clear evidence that ''totemo'' triggers a CI rather than a presupposition (see pgs. 169-171 on speaker- and subject-oriented readings of embedded “totemo”).

The evidence that ''motto'' triggers a CI is similarly incomplete. The implication cannot be directly rejected (pg. 120). Sawada claims that the implication is not background information; hence, not a presupposition. His evidence for this comes from the failure of the implication to pass the ''Wait a minute'' test. But, failure to pass this test does not tell us that the implication under discussion is a CI.

Sawada argues that ''yoppodo'' triggers a CI as well. The implication cannot be directly rejected (pg. 137). Sawada’s evidence that this implication is a CI comes from the “contextual felicity constraint'' (pgs. 137-138). According to Sawada, the implication associated with ''yoppodo'' is not presupposed because it is not required to be contextually entailed. He states (pg. 137, fn. 6) that Tonhauser et al. (2013) call this the ''strong contextual felicity constraint.'' But Tonhauser et al. (2013: 79-80) present evidence from GuaranÌ and English that certain expressions that are traditionally classified as presupposition triggers (e.g., possessive noun phrases) violate this constraint. Failure to respect this constraint does not provide conclusive evidence for CI-status.

Summing up, Sawada explores an impressive range of scalar modifiers that he claims trigger a CI. But the diagnostics for not-at-issue content are typically applied inconsistently and inconclusively.

It’s possible that the implications Sawada investigates do not form a unified kind. Potts (2005) argued both expressives and appositives contribute CIs. Subsequent research has identified differences between these two classes of expressions (Potts 2012). Similarly, so-called “presupposition triggers” are not a unified kind (Soames 1989 [2009], Karttunen 2016; Tonhauser et al. 2013; Tonhauser et al. 2018). “[W]e should take care not to overstate the unity [of expressions that appear to display multidimensionality - BZC]; many differences emerge, suggesting that we need to study each item on its own terms'' (Potts 2012).

CONCLUSION

Sawada's book is a solid contribution to the literature. Researchers working on (not-)at-issueness will find much to engage with. The book will be useful for further work on Japanese scalar modifiers. Importantly, this investigation of scalar modifiers provides a case study of the pitfalls that characterize attempts to test for not-at-issue content.

REFERENCES

Amaral, Patricia, Craige Roberts, & E. Allyn Smith. 2007. Review of The Logic of Conventional Implicatures by Chris Potts. Linguistics and Philosophy 30(6). 707-749.

AnderBois, Scott, Adrian Brasoveanu, & Robert Henderson. 2015. At-issue Proposals and Appositive Impositions in Discourse. Journal of Semantics 32. 93-138.

Bach, Kent. 1999. The Myth of Conventional Implicature. Linguistics and Philosophy 22(4). 327-366.

von Fintel, Kai. 2004. Would you believe it? The King of France is back! (Presuppositions and truth-value intuitions). In Anne Bezuidenhout & Marga Reimer (eds.), 315-341. Descriptions and Beyond. Oxford: Oxford University Press.

Heim, Irene. 1992. Presupposition Projection and the Semantics of Attitude Verbs. Journal of Semantics. 9. 183-221.

Karttunen, Lauri. 2016. Presupposition: What went wrong? In Mary Moroney, Carol-Rose Little, Jacob Collard, & Dan Burgdorf (eds.), 705-731. Proceedings of the 26th Semantics and Linguistic Theory Conference. Linguistic Society of America & Cornell Linguistics Circle.

Matthewson, Lisa, Hotze Rullmann, & Henry Davis 2007. Evidentials as epistemic modals. The Linguistic Variation Yearbook 7. 201-254.

McCready, Eric. 2010. Varieties of conventional implicature. Semantics & Pragmatics. 3(8). 1-57.

Murray, Sarah E. 2014. Varieties of update. Semantics & Pragmatics 7(2). 1-53.

Potts, Christopher. 2005. The Logic of Conventional Implicatures. Oxford: Oxford University Press.

Potts, Christopher. 2008. Wait a Minute! What Kind of Discourse Strategy is This? Annotated data. http://christopherpotts.net/ling/data/waitaminute/ (19 November 2018)

Potts, Christopher. 2012. Conventional implicature and expressive content. In Claudia Maienborn, Klaus von Heusinger, & Paul Portner (eds.), 2516-2536. Semantics: An International Handbook of Natural Language Meaning, Volume 3. Berlin: Mouton de Gruyter.

Shanon, Benny. 1976. On the two kinds of presupposition in natural language. Foundations of Language 14(2). 247-249.

Simons, Mandy, Judith Tonhauser, David Beaver, & Craige Roberts. 2010. What Projects and Why. In Nan Li & David Lutz (eds.), 309-327. Proceedings of the 20th Semantics and Linguistic Theory Conference. Linguistic Society of America & Cornell Linguistics Circle.

Snider, Todd. 2017. At-issueness [does not equal] anaphoric availability. In Patrick Farrell (ed.), 1-15. Proceedings of the Annual Meeting of the Linguistic Society of America (LSA) 2(39). Linguistic Society of America.

Soames, Scott. 1989 [2009]. Presupposition. In Dov Gabbay & Franz Guenthner (eds.), 553-616. Handbook of Philosophical Logic, vol. 4. Dordrecht: Reidel. Reprinted in Scott Soames 73-130. Philosophical Essays, vol. 1. Princeton, NJ: Princeton University Press, 2009.

Syrett, Kristen & Todor Koev. 2015. Experimental Evidence for the Truth Conditional Contribution and Shifting Information Status of Appositives. Journal of Semantics 32(3). 525-577.

Tonhauser, Judith, David I. Beaver, & Judith Degen. 2018. How projective is projective content? Journal of Semantics 35(3). 495-542.

Tonhauser, Judith, David I. Beaver, Craige Roberts, & Mandy Simons. 2013. Toward a taxonomy of projective content. Language 89(1). 66-109.


ABOUT THE REVIEWER

Brady Clark is a College Adviser and Associate Professor of Instruction at Northwestern University. He received his Ph.D. in linguistics in 2004 from Stanford University. His primary research areas currently are semantics and pragmatics, with specific interests in (not-)at-issueness, presupposition, definite descriptions, and metasemantics. He has published work on language change, focus, game-theoretic approaches to communication, and theories of language evolution.



Page Updated: 11-Jun-2019