We introduce several senses of the principle ofcompositionality. We illustrate the difference between them with thehelp of some recent results obtained by Cameron and Hodges oncompositional semantics for languages of imperfect information.

A semantic theory is committed to semantic monism just in case every particular semantic property posited by the theory is a member of the same kind. The commitment to semantic monism appears to draw some support from the need to provide a compositional semantics, since taking a single kind of semantic property as key to a semantic theory affords a uniform pattern on the basis of which the meaning of any given sentence can be compositionally determined. This line of support (...) highlights a tension between the principle of compositionality and taking a more pluralistic approach to semantic theorizing. In this essay, that tension is relieved, leaving open the prospect of a pluralistic semantic theory that observes the principle of compositionality. (shrink)

Charles S. Peirce’s pragmatist theory of logic teaches us to take the context of utterances as an indispensable logical notion without which there is no meaning. This is not a spat against compositionality per se , since it is possible to posit extra arguments to the meaning function that composes complex meaning. However, that method would be inappropriate for a realistic notion of the meaning of assertions. To accomplish a realistic notion of meaning (as opposed e.g. to algebraic meaning), Sperber (...) and Wilson’s Relevance Theory (RT) may be applied in the spirit of Peirce’s Pragmatic Maxim (PM): the weighing of information depends on (i) the practical consequences of accommodating the chosen piece of information introduced in communication, and (ii) what will ensue in actually using that piece in further cycles of discourse. Peirce’s unpublished papers suggest a relevance-like approach to meaning. Contextual features influenced his logic of Existential Graphs (EG). Arguments are presented pro and con the view in which EGs endorse non-compositionality of meaning. (shrink)

The Principle of Semantic Compositionality (sometimes called Frege''s Principle) is the principle that the meaning of a (syntactically complex) whole is a function only of the meanings of its (syntactic) parts together with the manner in which these parts were combined. This principle has been extremely influential throughout the history of formal semantics; it has had a tremendous impact upon modern linguistics ever since Montague Grammars became known; and it has more recently shown up as a guiding principle for a (...) certain direction in cognitive science.Despite the fact that The Principle is vague or underspecified at a number of points — such as what meaning is, what counts as a part, what counts as a syntactic complex, what counts as combination — this has not stopped some people from viewing The Principle as obviously true, true almost by definition. And it has not stopped other people from viewing The Principle as false, almost pernicious in its effect. And some of these latter theorists think that it is an empirically false principle while others think of it as a methodologically wrong-headed way to proceed. (shrink)

Starting from the familiar observation that no straightforward treatment of pure quotation can be compositional in the standard (homomorphism) sense, we introduce general compositionality, which can be described as compositionality that takes linguistic context into account. A formal notion of linguistic context type is developed, allowing the context type of a complex expression to be distinct from those of its constituents. We formulate natural conditions under which an ordinary meaning assignment can be non-trivially extended to one that is sensitive to (...) context types and satisfies general compositionality. As our main example we work out a Fregean treatment of pure quotation, but we also indicate that the method applies to other kinds of context, e.g. intensional contexts. (shrink)

This is the second part of a two-part article on compositionality, i.e. the principle that the meaning of a complex expression is determined by the meanings of its parts and the way they are put together. In the first, Pagin and Westerståhl (2010), we provide a general historical background, a formal framework, definitions, and a survey of variants of compositionality. It will be referred to as Part I. Here we discuss arguments for and against the claim that natural languages have (...) a compositional semantics. We also discuss some problem cases, including belief reports, quotation, idioms, and ambiguity. (shrink)

I argue that compositionality (in the sense of homomorphic interpretation) is compatible with radical and pervasive contextual effects on interpretation. Apparent problems with this claim lose their force if we are careful in distinguishing the question of how a grammar assigns interpretations from the question of how people figure out which interpretations the grammar assigns. I demonstrate, using a simple example, that this latter task must sometimes be done not by computing a derivation defined directly by the grammar, but through (...) the use of pragmatic background knowledge and extragrammatical reasoning, even when the grammar is designed to be fully compositional. The fact that people must sometimes use global pragmatic mechanisms to identify truth conditions therefore tells us nothing about whether the grammar assigns truth conditions compositionally. Compositional interpretation (or the lack thereof) is identifiable not by the mechanisms necessary to calculating truth conditions, but by the structural relation between the interpretation of a phrase in context and the interpretations of its parts in that same context. Even if this relation varies by context, an invariant grammar is possible if grammars can “invoke” pragmatic concepts; but this does not imply that grammatical theory must explain these concepts or incorporate a theory of pragmatics. (shrink)

Typical accounts of imagistic content have focused on the apparent analog character or continuous variability of images. In contrast, I consider the distinctive features of digital images, those composed of finite sets of discrete pixels. A rich source of evidence on digital imagistic content is found in the content-preserving algorithms that resize and reproduce digital images on computer screens and printers. I argue that these algorithms reveal a distinctive structural feature: digital images are always compositional (their parts contribute systematically to (...) overall content), but never inverse compositional (atomic parts may be replaced nonsynonymously without changing content). This indicates a sharp contrast with linguistic representations, which may or may not be compositional, and may or may not be inverse compositional. I argue this result sheds new light on the claim that imagistic content is inherently perspectival. (shrink)

This paper proposes that the meanings of some natural language expressions should be thought of as functions on their own continuations. Continuations are a well-established analytic tool in the theory of programming language semantics; in brief, a continuation is the entire default future of a computation. I show how a continuation-based grammar can unify several aspects of natural language quantification in a new way: merely stating the truth conditions for quantificational expressions in terms of continuations automatically accounts for scope displacement (...) and scope ambiguity. To prove this claim, I exhibit a simple finite context-free grammar with a strictly compositional semantics in which quantificational NPs are interpreted in situ but take semantic scope over larger constituents. There is no Quantifier Raising (nor any use of a level of Logical Form distinct from overt syntax), no Cooper Storage (or similar mechanisms used in many recent HPSG, Categorial, or Type-logical treatments), and no need for type-shifting (as in Hendriks' Flexible Types account). Continuations also provide a natural account of generalized coordination that does not require either type-shifting or type polymorphism. Compositionality issues are discussed in some detail. (shrink)

The success of deep learning in natural language processing raises intriguing questions about the nature of linguistic meaning and ways in which it can be processed by natural and artificial systems. One such question has to do with subword segmentation algorithms widely employed in language modeling, machine translation, and other tasks since 2016. These algorithms often cut words into semantically opaque pieces, such as ‘period’, ‘on’, ‘t’, and ‘ist’ in ‘period|on|t|ist’. The system then represents the resulting segments in a dense (...) vector space, which is expected to model grammatical relations among them. This representation may in turn be used to map ‘period|on|t|ist’ (English) to ‘par|od|ont|iste’ (French). Thus, instead of being modeled at the lexical level, translation is reformulated more generally as the task of learning the best bilingual mapping between the sequences of subword segments of two languages; and sometimes even between pure character sequences: ‘p|e|r|i|o|d|o|n|t|i|s|t’ → ‘p|a|r|o|d|o|n|t|i|s|t|e’. Such subword segmentations and alignments are at work in highly efficient end-to-end machine translation systems, despite their allegedly opaque nature. The computational value of such processes is unquestionable. But do they have any linguistic or philosophical plausibility? I attempt to cast light on this question by reviewing the relevant details of the subword segmentation algorithms and by relating them to important philosophical and linguistic debates, in the spirit of making artificial intelligence more transparent and explainable. (shrink)

Compositionality has been a central concept in linguistics and philosophy for decades, and it is increasingly prominent in many other areas of cognitive science. Its status, however, remains contentious. Here, I reassess the nature and scope of the principle of compositionality (Partee, 1995) from the perspective of psycholinguistics and cognitive neuroscience. First, I review classic arguments for compositionality and conclude that they fail to establish compositionality as a property of human language. Next, I state a new competence argument, acknowledging the (...) fact that any competent user of a language L can assign to most expressions in L at least one meaning which is a functiononlyof the meanings of the expression's parts and of its syntactic structure. I then discuss selected results from cognitive neuroscience, indicating that the human brain possesses the processing capacities presupposed by the competence argument. Finally, I outline a language processing architecture consistent with the neuroscience results, where semantic representations may be generated by a syntax‐driven streamandby an “asyntactic” processing stream, jointly or independently. Compositionality is viewed as a constraint on computation in the former stream only. (shrink)

In this paper, we defend a traditional approach to semantics, that holds that the outputs of compositional semantics are propositional, i.e. truth conditions. Though traditional, this view has been challenged on a number of fronts over the years. Since classic work of Lewis, arguments have been offered which purport to show that semantic composition requires values that are relativized, e.g. to times, or other parameters that render them no longer propositional. Focusing in recent variants of these arguments involving quantification and (...) binding, we argue that a correct understanding of how composition works gives no reason to relativize semantic values, and that propositional semantic values are in fact the preferred option. We take our argument to be mainly empirical, but along the way, we defend some more general theses. Simple propositional semantic values are viable in composition, we maintain, because composition is itself a complex phenomenon, involving multiple modes of composition. Furthermore, some composition principles make adjustments to the meanings of constituents in the course of composition. These adjustments are by triggered syntactic environments. We argue such small contributions of meaning from syntactic structure are acceptable. (shrink)

A remarkable thing about natural language is that we can use it to share our beliefs and thoughts about the world with other speakers of our language. In cases of successful communication, beliefs seem to be transferred from speakers to hearers by means of the hearer recovering the contents of the speaker’s utterances. This is so natural to us that we take it for granted in our everyday life, and rarely stop to think about how it's is possible. Nevertheless, it's (...) a phenomenon that calls for explanation. It is natural to expect that natural language semantics has a key explanatory role to play here. In order to understand this role, we must relate the semantic values assigned to sentences by semantic theory with the content of our speech acts. The simplest possible relation would be identity; the meaning of a sentence is simply the belief expressed by an assertion of the sentence in a given context of utterance. However, a number of problem cases in the literature suggest that this cannot be the case. This dissertation offers a critical assessment of the arguments for distinguishing the semantic value of a sentence from it's so called assertoric content, focusing on problems arising from the analysis of tense and temporal expressions. I conclude that they are indeed distinct, and offer a constructive account of how they must be related in order to allow for an explanation of communicative success. (shrink)

When reasons are given for compositionality, the arguments usually purport to establish compositionality in an almost a priori manner. I will rehearse these arguments why one could think that compositionality is a priori true, or almost a priori true, and will find all of them inconclusive. This, in itself, is no reason against compositionality, but a reason to try to establish or defend the principle on other than quasi-a priori grounds.

Recent attempts to define and support realism in semantics seem to acknowledge, as the only defence from skeptical attacks to the notion of meaning, a flat acceptance of the existence of representational relations between language and things in the world. In this paper I reconsider part of the mistrust about the normative character of meaning, in order to show that some of the worries urging the realists to cling on representationalism actually rest on misconceptions. To the contrary, I suggest that (...) normativity is the main strength of a stable realist stance in semantics. Support to this suggestion comes from the reanalysis of some oft-ignored sellarsian themes. (shrink)