|
Radical contextualism and open compositionality: a metatheoretical proposal
Contextualismo radical y composicionalidad abierta: una propuesta metateórica
Pablo David Chávez Carvajal Universidad Nacional Autónoma de México, México https://orcid.org/0000-0002-1457-4705
|
|
|
INFORMACIÓN DEL ARTÍCULO |
ABSTRACT/RESUMEN |
|
Recibido el: 15/7/2025 Aceptado el: 31/8/2025
Keywords: Philosophy of language, semantics, radical contextualism, compositionality, open compositionality
Palabras clave: Filosofía del lenguaje, semántica, contextualismo, composicionalidad, composicionalidad abierta |
Abstract: In this paper, I explore the relationship between radical contextualism and compositionality. Radical contextualism is a family of theories defending that a sentence’s meaning in natural language is always invariant because it is too broad and unspecified to be otherwise and requires contextual adjustment. Compositionality is the idea that a sentence’s complex meaning comprises its parts. I shall evaluate whether radical contextualism benefits from including compositionality to explain linguistic meaning in natural language. Including compositionality might allow us to formalize aspects of radical contextualism and explain meaning-formation more precisely. I shall argue, nonetheless, that the classical notion of composition fails to account for the interpretation of sentence meaning in natural languages (as guided by a radical contextualist theory). An open compositionality scheme is crucial since I understand meaning-determination as a decision-making task. To explain how we understand sentences in natural language, open compositionality must be used to formulate a radical contextualist theory. I shall also offer a novel metatheory encompassing open compositionality and radical contextualism.
Resumen: En este artículo, exploro la relación entre el contextualismo radical y la composicionalidad. El contextualismo radical es un conjunto de teorías que defienden que el significado de una oración en el lenguaje natural es siempre invariante porque es demasiado amplio y poco específico como para ser de otra manera y requiere un ajuste contextual. La composicionalidad es la idea de que el significado complejo de una oración se compone de sus partes. Evaluaré si el contextualismo radical se beneficia de la inclusión de la composicionalidad para explicar el significado lingüístico en el lenguaje natural. La inclusión de la composicionalidad podría permitirnos formalizar aspectos del contextualismo radical y explicar la formación del significado con mayor precisión. No obstante, argumentaré que la noción clásica de composición no logra explicar la interpretación del significado de las oraciones en los lenguajes naturales (tal y como lo indica una teoría contextualista radical). Un esquema de composicionalidad abierta es crucial, ya que entiendo la determinación del significado como una tarea de toma de decisiones. Para explicar cómo entendemos las oraciones en el lenguaje natural, se debe utilizar la composicionalidad abierta para formular una teoría contextualista radical. También ofreceré una nueva metateoría que abarca la composicionalidad abierta y el contextualismo radical. |
Introduction
Roughly stated, the principle of compositionality (commonly attributed to Frege)1 states that the meaning of a complex expression is determined by its constituents: for example, the complex meaning of a sentence is determined by its words and their structure. Compositionality has been used to explain how the meaning of an expression in a natural language is composed. Modeling language users’ linguistic competence (i.e., accounting for how they understand sentence meanings) putatively explains how language users can learn natural languages and, in turn, how they understand sentences they have not heard before. Several scholars have, however, argued that radical contextualism is largely incompatible with compositionality.
Radical contextualism in the philosophy of language is a family of theories that more or less follow Recanati (2010) in defending the following thesis: “[I]n general (i.e. not only in the special case of indexicals), the propositional contribution of an expression is not fully determined by the invariant meaning conventionally associated with the expression type but depends upon the context”. (p. 17)
This view can be divided into moderate and radical contextualism.
· Moderate contextualism assumes that the proposition a sentence conveys is semantically complete and its truth value is determined by context-sensitive aspects of the sentence’s use.
· In radical contextualism, the invariant meaning is always unspecified, too wide to be made specific, and in need of contextual adjustment. This suggests a high degree of contextual sensitivity.
I shall concentrate exclusively on radical contextualism. The difficulty in reconciling it with compositionality lies in (a) how grammar allows the assignment of a complex sentence’s interpretation and (b) how language users understand complex sentences, guided only by their context (the time and place of utterance).
Consider the sentence:
We have a lame duck situation [1],
uttered in three different contexts:
1. In a park. It would then have a conventional meaning, namely that there is a lame duck and it probably needs help.
2. In a business environment in the United States. Here, the situation is that a manager is about to step down.
3. In a cricket match. A batter has been dismissed without making runs, and this has happened uninterestingly.
The first context is the only one that can be interpreted purely through grammar. The other two interpretations require contextual adjustment.
The following questions thus arise: can the notion of compositionality be preserved in a radical contextualist theory? Would compositionality’s inclusion in radical contextualism be of any use? Do we need a notion of compositionality that differs from classical compositionality? To answer these questions, I shall first critically describe what compositionality is, some of its applications and variations, and the most common objections. I shall then discuss the relationship between compositionality and contextualism and then characterize contextualism in general to highlight the difficulties this relationship faces. Next, I shall formulate a possible radical-contextualist response—one that considers Recanati’s (2010) proposal to weaken classical compositionality. However, such a reply fails because it does not resolve the salient variations that can appear when defending the contextualist thesis—mainly that a sentence’s meaning always requires contextual adjustment.
However, classical compositionality suggests that the level of meaning which we should consider compositional is not radically context sensitive. According to classical compositionality, this level should be sufficient; to maintain radical contextualism motivated by contextual sensitivity, we should therefore adopt a theory of open compositionality, as García-Ramírez (2019) suggests. This allows for the inclusion of significant variability in linguistic meaning, along with compositionality. I shall explain this notion and argue that open compositionality is the only way for radical contextualists to preserve the advantages of compositionality. The discussion presupposes radical contextualism, but this is a family of theories that can differ from each other, so I shall present a suitable metatheory. This entails developing a set of propositions expressing the conditions a radical contextualist theory must fulfill to maintain compositionality.
What is compositionality and how useful has it been in the philosophy of language?
![[2]](7.%20Radical%20contextualism%20and%20open%20compositionality.%20A%20metatheoretical%20proposal-Pablo%20David%20Chávez%20Carvajal%20_archivos/image001.png){kind=small}
Language users can produce sentences they have
never written or verbalized before. These sentences can, in turn, be understood
by other language users who have not previously encountered them. For example,
the following sentence (or one equivalent in meaning) has likely never been
produced before:
A young rugby player and his uncle are playing bridge in a purple
house in Montevideo, Minnesota.
Most English speakers will have no difficulty in understanding this sentence, even if they have never heard or read it before, since they can identify the parts of the sentence. This indicates that the ability to understand a natural language’s sentences lies in how their parts cohere to generate distinct meanings. The number of complex expressions language users can produce and understand seems to be unlimited. But how can limited beings like us be so linguistically productive? How do we learn any number of languages, and the unlimited number of meaningful expressions that can be formed from each language’s more or less stable and restricted vocabulary? Such a phenomenon requires explanation, and the best hypothesis involves natural languages’ compositionality. Here is how Elbourne (2011) characterizes the principle of compositionality: “The meaning of a complex phrase is determined solely by the meaning of its parts and their syntactic arrangement” (p. 28). (See Zimmerman, 2020, for a contemporary reconstruction of possible Fregean composition) This principle (or one of its variations) has been used to explain competent speakers’ linguistic ability, which (hypothetically) allows such speakers of a natural language to understand an infinite set of linguistic expressions, based on their knowledge of specific, elementary categories.2 This is called “productivity.” As human beings with cognitive limitations, we cannot know an infinite number of meanings immediately. Elbourne explains that meaning is systematic because it allows complex expressions to be related. His principle thus establishes a methodological claim about how we should study meaning in natural language.
Nevertheless, Jaszczolt (2018) has noted the following: “Compositionality is not only a methodological claim but also an epistemological and metaphysical one, based on the argument of the productivity and systematicity of patterns of interaction in conversation”. (p. 80)
Compositionality is epistemological because it lets us account for how speakers know meanings in natural language. This starts with their knowledge of languages’ essential elements, which cannot be explained in more elementary terms. Szabó (2012) expresses the idea more precisely:
The argument from systematicity states that anyone who understands a number of complex expressions e1..., en understands all other complex expressions that can be built up from the constituents of e1..., en using syntactic rules employed in building up their structures. Since this is so, there must be something competent speakers know (perhaps tacitly) based on which they can determine what the complex expressions built through such recombination mean. (p. 77)
Compositionality is metaphysical because it determines what kind of composition natural language sentences must have if they are to constitute meaningful expressions. An unordered juxtaposition of strings of characters implies a meaningless concatenation of marks and sounds. A truly meaningless strings of characters would be something like “Zdfbdf qrte szgwet” (i.e., completely random letters, assuming randomness is meaningless).
That said, there are concrete semantic theories that use compositionality. Examples include Montague’s semantics (Dowty et al., 1981) and Davidson’s (2001) truth-conditional semantics. (For a detailed study of this type of semantics, see Lepore & Ludwig, 2009). However, these are semantic theories in the formal sense—theories that are shaped by the following properties:
A compositional theory of meaning for a language L is a formal theory that enables anyone who understands the language in which the theory is formulated to understand the primitive expressions of L, and the complex expressions of L based on the understanding of the primitives. (Lepore & Ludwig, 2009, p. 18)
Considering the characteristics usually attributed to compositionality, we can add to this a fundamental property of natural languages: learnability. The arguments used to defend such notions are therefore referred to as learnability arguments. (This property is most notably defended by Davidson, 2001) Here is a typical one, in which L stands for any natural language (see Pagin, 2012, p. 514, where the argument comes from):
1. There is a sentence in L, in any context c, for each of the infinite number of existing meanings.
2. We can learn L (and thus understand all these sentences that convey each of the infinite number of meanings).
3. The mechanism of directly learning sentence meanings allows us to learn only a finite number of such meanings.
4. We can learn L, and understand expressions for the infinite number of meanings, only if we can follow compositional rules for constructing the infinite number of expressions in L, based on the finite number of expressions that can be directly learned (using the mechanism mentioned in Premise 3). In other words, we can learn L only if L has compositional semantics.
5. L has compositional semantics.
However, for Pagin (2012), the learnability argument3 leads to a problem, which arises from the assumption that the infinity of sentences speakers never use are meaningful. This problem is generated by one of the argument’s premises and a presupposition. In Premise 1, it is assumed that a natural language, such as English, has an infinity of meanings. At the same time, it is presupposed that humans cannot know an infinite number of meanings; rather, there are infinite meanings which humans can potentially learn. If so, English will have an infinite set of meanings that speakers do not use, leading to the production of linguistic meanings that seem unnecessary for communication. The best justification for holding that any natural language has infinite meanings is that language semantics is compositional: these unused sentences’ meanings can be constructed, even though they are not used. However, if this is the justification for holding P1 (that there is an infinity of meanings we never use) then one commits a petitio principii: one presupposes compositionality rather than proving its necessity.
This demonstrates how undesirable consequences arise when we explain meaning composition as part of an attempt to formulate a compositional theory. We can see that “getting to the bottom of the composition of meaning is not a simple task. There are also difficulties involved in combining the meaning of adjectives and nouns” (Jaszczolt, 2023, p. 65). For instance, a falsified Picasso is, strictly speaking, not a Picasso. Since such difficulties arise when appealing to compositionality in natural language, one can ask: What results from contextual sensitivity’s inclusion in the meaning of natural language sentences? By contextual sensitivity, I mean the following: A sentence s will have the property of contextual sensitivity (cs) if and only if how s’s meaning is understood changes depending on the occurrence of s’s use, which is determined by the contextual characteristics (time and place) of s’s use.
Given how context can affect a sentence’s meaning, speakers’ context and intention when uttering a sentence can determine that meaning only broadly.
Consider sentence [1] again. Plausibly, we can determine which of its three meanings applies only when we recognize the speaker’s communicative intention. It would, therefore, be challenging to find compatibility between contextual sensitivity and compositionality. My discussion will focus on the relationship between compositionality and radical contextualism—a thesis motivated by contextual sensitivity—because several scholars have argued that the two theses are incompatible.4
The theoretical incompatibility between compositionality and radical contextualism
How can meaning in natural language be studied if there is context sensitivity? How context determines sentence meaning is controversial. A context’s representational accuracy depends on the contextualist theory postulating what role it plays or what it is. As Ciecierski & Grabarczyk (2020) have written:
It is a truism that context, content, and meaning interact. On the one hand, context determines what is communicated in the discourse; on the other hand, what is communicated in the discourse shapes the situation in which the discourse is embedded. Philosophers who agree about these basic facts disagree about nearly everything else. (p. 1)
We can characterize context thus: “We call these things ‘contexts’ because they are being used to represent the concrete situation in which a sentence event takes place” (Stalnaker 2018, p. 105). We can define radical contextualism as follows:
[In radical contextualism, the] invariant meaning of a non-indexical expression type is too unspecified and/or too rich to render a literal interpretation of its token in any context. Consequently, contextual adjustment is required in every context to determine its propositional contribution. (Huang, 2017, p. 964)5
However, if we require contextual adjustment then we also require more than an elementary notion of context (i.e., time and place). Plausibly, we can characterize context6 in a very general way, compatible with my purposes here, to present a general radical-contextualist thesis. Pollock’s characterization may be useful in this regard. She writes:
Context here should be understood in the broad sense, to include things like interlocutors’ background knowledge as well as physical features of the environment of the speech exchange. (Pollock, 2020, p. 249)7
As such, we can understand what is required to clarify the notion of radical contextualism as follows: Consider any natural language sentence S that requires contextual adjustment to fix its meaning M1 at the time of utterance. Such an adjustment will be achieved if and only if either the interlocutor’s background knowledge or the physical characteristics of the environment in which S is used determine what M1 means at a particular time.
For example, as the Introduction explained, contextual adjustment is required for sentence [1]:
· This expression might be used in a business environment in the United States when a manager is about to step down. The interlocutor recognizes the meaning of the idiomatic expression “lame duck” (to express that this manager is about to leave), which represents a situation of interest to the people having the conversation.
· The expression might also be used in a cricket context when a batter has been dismissed without making runs and nothing makes this situation interesting. Recognizing the cricket jargon, the interlocutor knows that “duck” means that the batsman has not scored a single run. Given the circumstances of the match, however, this occurrence could have been more interesting.
We can therefore describe why radical contextualism is incompatible with compositionality if there must always be a contextual fit based on a sentence’s interpretation (when considering radical contextualism). How can we maintain the notion that the meaning of a complex sentence is determined solely by the meanings of its parts and their syntactic arrangement? Compositionality would hold only in the sentence’s grammar, so contextual adjustment could not require compositionality. This is because compositionality does not depend on factors that are non-linguistic, such as background knowledge and the environment of a sentence. Is there a way to maintain both radical compositionality and contextualism? One way involves postulating a series of rules that allow us to invoke the need for contextual adjustment when explaining the variability in a sentence’s linguistic meanings. As we have seen, this weakens compositionality.
To illustrate this, we can recognize the following context-sensitive lexical rule (1*), as formulated by Recanati (2010):
I(a)c = f(c)
The characterization is as follows: the conventional meaning of a sentence s determines a function f, which, given a context c, momentarily alters its meaning (Recanati, 2010, 37).8 More precisely, this rule models how a sentence’s contextual adjustment occurs (as discussed in Section 1). A sentence’s conventional meaning becomes a function of the context. In other words, the conventional meaning is adjusted to fit the context of the sentence. This provides a way to reconcile (i) what is grammatically determined in a sentence and (ii) the context sensitivity that can cause that sentence’s meaning to vary.
Let us return to [1] in its second meaning. The conventional meaning of “lame duck” appears as a contextually determined function in the business environment. However, other types of circumstances are excluded (e.g., that there is a mallard duck present or a cricket match going on). Having a theory that preserves compositionality (even while accepting that natural language displays contextual sensitivity) is highly desirable because Recanati’s proposal produces rules that establish a procedure under which natural language sentences obtain various meanings.9
Briciu (2020) has, however, objected that including rules that adjust compositionality to fit extreme contextual sensitivity does not work. In his view, clear and stable contextual parameters cannot fix this as a type of sensitivity, for at least two reasons:
1. The meaning of a sentence in natural language depends on many factors and can vary widely.
2. There is no well-determined limit on the number of pragmatic factors the emission context can influence.
If this is the case, then it seems to be impossible to have a rule for weak compositionality. One would not always be able to predict which types of pragmatic phenomena will influence a sentence’s content.
Briciu has raised such a concern from a different perspective: “If natural languages are compositionally weak and their expressions are radically context-sensitive, a single syntactic structure will contribute in more than one way to the interpretation of complex expressions”. (2020, 217)
Consider the sentence:
John observed a man using his computer [3].
Here, we would usually say that there is a single syntactic structure with at least two interpretations: (i) John observed a man through his computer or (ii) a man was using John’s computer. This shows that the algorithms which a weak compositionality rule obtains can hardly accommodate the degree of variability, owing both to extreme contextual sensitivity and to the sheer diversity of meanings that can be adjusted via pragmatic modulation. As Recanati suggests, weakening compositionality therefore does not actually have the benefits it hypothetically might have had. This makes it very complicated to retain compositionality in a theory that explains communication by accepting the breadth of linguistic meanings in natural language, so weakened compositionality fails to bring the benefits of compositionality to theories that study linguistic meaning in natural language. If weakening compositionality does not allow us to reconcile weakened and non-weakened compositionality, then there seems to be a problem with modifying our understanding of compositionality.
That said, one can argue that compositionality is valid and can be maintained, even if sentence meanings can be determined only subject to contextual dependence. Open compositionality makes it possible to explain how the meaning of a complex sentence is formed from its components (which belong to various domains besides the linguistic), so we can include a high degree of context sensitivity when examining a natural language sentence’s meaning.
What is open compositionality?
When characterizing open-ended compositionality, one must take into account a thesis about the nature of language that García-Ramírez (2019) calls the “Lewisian Commitment”10: “Natural languages are, first and foremost, things that can be learned, developed and used by human beings given the limits and nature of their cognitive resources” (2019, p. 92). At first sight, such a commitment appears compatible with the formulation of compositionality explicated above. This is because open compositionality seeks to account for the characteristics that allow a language to be learned and its complex expressions to have meaning.
However, García-Ramírez has argued that the knowledge through which we learn a language stems from various epistemic dimensions11: not only those dependent on the language itself (as is the case with compositionality), but also cultural aspects, environmental stimuli, and emotional elements, among other things. What matters most when it comes to interpreting sentences is that, in open compositionality, we seek the most plausible interpretation of a complex expression in natural language. In other words, there are two requirements when interpreting a sentence’s meaning: it must (i) be compatible with contextual variability and (ii) require little cognitive effort to interpret.
García-Ramírez states: “According to this view, compositional processes are only one among multiple different procedures required to account for the meaning of complex expressions” (2019, 16). There may be other ways, without compositional processes, to describe how complex sentences in natural languages have meaning. From a methodological point of view, open compositionality is a decision-making process in which we interpret meaning as economically as we can (i.e., based on speakers’ informational constraints). Open compositionality also accepts contextual sensitivity because a sentence can convey different meanings depending on its context (as discussed above).
Given all this, compositionality (as formulated here) fails to fulfill the Lewisian Commitment. No single source determines anyone’s knowledge of language. This definition of open-ended compositionality can be stated more rigorously as follows: Depending on contextual demands, the procedure of interpreting a sentence may sometimes involve heuristic strategies, syntactic algorithms, or both (either in parallel or as mutual backup strategies).
From this perspective, the context of a sentence12 is what determines the meaning of a complex expression. Let us return to [1] in its three occurrences:
1. In a park. It would have a conventional meaning, namely that there is a lame duck and it probably needs help.
In this case, the complex meaning is determined by the conventional meanings of the words and by information about the world (the physical environment).
2. In a business environment in the United States. Here, the situation is that a manager is about to step down.
In this case, the complex meaning is determined by semantic knowledge of idiomatic expressions related to business practices in American English. An important circumstance also merits the use of “lame duck.”
3. In a cricket match. A batter has been dismissed without making runs, and this has happened uninterestingly.
In this case, the complex meaning is determined by knowledge of cricketing jargon and by the identification of an event that merits the adjective “lame.”
According to the notion of open-ended compositionality, these complex meanings are shaped by the context, background knowledge, and physical environment in which a given sentence is used. What helps us to understand the meanings of each formulation is that “lame duck” is a polysemous expression. Its meaning can come from different sources, so identifying them helps us to discern which of its possible meanings it receives from them. By letting us clarify the type of heuristic strategies we use to interpret statements, this example also illustrates a criterion that determines how we make decisions when interpreting a sentence in natural language. We can find the most economical interpretation of each occurrence of [1] because open compositionality appeals to a criterion that lets us discern the processes for disambiguating an expression in natural language.
Open compositionality admits that (most of the time) a complex sentence’s linguistic meaning cannot be determined solely by the meanings and syntactic arrangement of its parts. It requires many other processes, which may be compatible with context sensitivity. Now, we are seeking a contextualist theory that considers such sensitivity while accounting for complex sentences’ configurations of meaning, so we should appeal to some form of open compositionality. If radical contextualist theories are based on open compositionality, then they can overcome the problems classic compositionality faces, as I showed above. This allows me to formulate the following metatheoretical claim:
A radical contextualist theory can include compositionality if and only if that theory is based on a notion of sentence meaning that is configured according to open compositionality.
In this case sentence meaning is the most cognitively economical interpretation of a sentence in everyday discourse that is formulated at some specific moment in time and determined by some decision-making process.
Conclusions
I have presented a critical description of compositionality’s formulation, some of its applications, some variations, and the most frequent objections. I have also discussed compositionality and radical contextualism. In so doing, I have provided a characterization of radical contextualism in general—one that may be compatible with various manifestations of radical contextualism. I employed a relevant notion of context to clarify this.
The need for compatibility between compositionality and radical contextualism involves the idea that a complex expression’s meaning comes from its parts and syntactic arrangement—taking contextual sensitivity into account makes this difficult. Recanati’s proposal to weaken compositionality by making it a function of sentence context fails because it attempts to cohere context-sensitive lexical items, which can hardly be treated formally. The main tension between classical compositionality and contextualism lies in contextual sensitivity. We should modify the notion of classical compositionality to endorse open compositionality when formulating radical contextualist theories.
Open compositionality implies that complex sentence meanings are determined by a series of processes—not necessarily linguistic ones—in which we choose the meaning that presents an interlocutor with the least cognitive effort. Such processes include the contextual variation to which sentence meanings are subject in any natural language, so compositionality can help a contextualist theory only if it is open compositionality. To characterize this, I have proposed a concomitant metatheoretical principle.
It will be useful to study this principle in future research, especially when one considers open compositionality’s relevance to radical contextualist theories that are based on the need for contextual adjustment.
Notes
[1] Despite this attribution, finding a canonical definition of the compositionality principle in Frege’s work is not easy. According to Janssen, “[t]he most well-known fragment of Frege which reminds of compositionality is the first sentence of Gedankegefüge [Logical Investigations] in the translation by Geach & Stoothoff [Geach & Stoothoff, 1977]: It is astonishing what language can do. With a few syllables, it can express an incalculable number of thoughts so that even a thought grasped by a terrestrial being for the very first time can be put into the form of words which will be understood by someone to whom the thought is entirely new. This would be impossible if we were not able to distinguish parts of the thoughts corresponding to the parts of a sentence so that the structure of the sentence serves as the image of the structure of the thoughts” (Janssen, 2001).
2 Given this paper’s metatheoretical nature, I shall limit myself to a very general analysis of compositionality. For a more detailed study, see Szabó (2012), who puts forward three compositionality theses in different disciplines: philosophy, linguistics, and psychology.
3 Pagin (2012) mentions a second problem. He argues that learnability (as the argument conceives it) does not require compositionality, only that languages be computable. This topic, however, is orthogonal to my discussion in this paper.
4 Peter Lasersohn (2012) discusses the reasons for this argument—an argument he does not find entirely convincing.
5 For more on radical contextualist theories, see Carston (2012) and Travis (2008).
6 David Lewis’ notion of context in compositional semantics has also been highly influential. He states that a “context is a location – time, place, and possible world – where an utterance is said” (Lewis,1997; see Stalnaker, 2018 for more).
7 I have chosen this notion because it is compatible with a general thesis underpinning radical contextualism and can include diverse theories.
8 I have adjusted this rule to fit the terminology used in this paper without considering the definitions Recanati uses in his theory.
9 Another way to characterize compositionality is as follows: maintain strict compositionality but separate it from the truth values of any interpretation that depends on how speakers judge the truth or falsity of an utterance. In this approach, we have compositional semantics that focuses on the syntactic and lexical mechanisms that formally represent meanings in a language. In other words, there is an abstraction of a particular language’s meaning —an abstraction that excludes speakers’ use of linguistic expressions.
10 David Lewis (1975) established two perspectives on human language. The first perspective is formal— a language is an abstract entity consisting of a set of marks and sounds, independent of their use. From the other perspective, human populations are guided by conventions of truthfulness and trust in using these marks and sounds. According to García- Ramírez, Lewis argued that human languages can be understood in purely formal terms. However, Lewis did not actually develop this commitment. Thus, one of the motivations behind open compositionality involves carefully configuring the Lewisian Commitment.
11 García-Ramírez appeals to empirical evidence to support his argument for open compositionality. However, in this paper I assume that García-Ramírez’s theoretical apparatus is plausible without carefully evaluating this evidence (see García-Ramírez, 2019, Ch. 3).
12 Understood in the sense proposed by Pollock (2020).
References
Briciu, A. (2020). Compositionality in Truth-Conditional Pragmatics. In T. Ciecierski and P. Grabarczyk (Eds.), The Architecture of Context and Context-Sensitivity (pp. 205–225). Springer.
Carston, R. (2012). Word Meaning and Concept Expressed. The Linguistic Review, 29(4), 607–623.
Ciecierski, T. & Grabarczyk, P. (2020). An Introduction: The Architecture of Context and Context-Sensitivity. In T.G. Ciecierski and P. Grabarczyk (Eds.), The Architecture of Context and Context-Sensitivity (pp. 205–225). Springer.
Davidson, D. (2001). Theories of Meaning and Learnable Languages. In D. Davidson, Inquiries into Truth and Interpretation (pp. 1–15). Oxford University Press.
Dowty, D., Wall, R. & Peters, S. (1981). Introduction to Montague Semantics, no. 11 in Synthese Language Library, Reidel.
Elbourne, P. (2011). Meaning: A Slim Guide to Semantics. Oxford University Press.
García-Ramírez, E. (2019). Open Compositionality: Toward a New Methodology of Language. Lexington Books.
Geach, P. T. & Stoothoff, R. H. (1977). Logical Investigations. Gottlob Frege. Basil Blackwell.
Huang, M. (2017). A Plea for Radical Contextualism. Synthese, 194(3), 963–988.
Janssen, T. M. V. (2001). Frege, Contextuality and Compositionality. Journal of Logic, Language and Information 10(1), 115–136.
Jaszczolt, K. (2018). Meaning in Linguistic Interaction: Semantics, Metasemantics, Philosophy of Language. Oxford University Press.
Jaszczolt, K. (2023). Semantics, Pragmatics, Philosophy: A Journey through Meaning. Cambridge University Press.
Lasersohn, P. (2012). Contextualism and Compositionality. Linguistics and Philosophy, 35(2), 171–189.
Lepore, E. & Ludwig, K. (2009). Donald Davidson’s Truth-Theoretic Semantics. Clarendon Press.
Lewis, D. (1975). Languages and Language. In K. Gunderson (Ed.), Language, Mind, and Knowledge (pp. 3–35). University of Minnesota Press.
Lewis D. (1997). Index, Context, and Content. In: Papers in Philosophical Logic. Cambridge Studies in Philosophy (pp. 21–44). Cambridge University Press.
Pagin, P. (2012). Communication and the Complexity of Semantics. In W. Hinzen, M. Werning, and E. Machery (Eds.), The Oxford Handbook of Compositionality (pp. 510–529). Oxford University Press.
Pollock, J. (2020). Occasion-Sensitivity and What Is Said. In T. Ciecierski and P. Grabarczyk (Eds.), The Architecture of Context and Context-Sensitivity (pp. 245–264). Springer.
Recanati, F. (2010). Truth-Conditional Pragmatics. Oxford University Press.
Stalnaker, R. (2018). David Lewis on Context. In D. Ball and B. Rabern (Eds.), The Science of Meaning: Essays on the Metatheory of Natural Language Semantics (pp. 98–112). Oxford University Press.
Szabó, Z. G. (2012). The Case for Compositionality. In W. Hinzen, E. Machery, and M. Werning (Eds.), The Oxford Handbook of Compositionality (pp.64 –80). Oxford University Press.
Travis, C. (2008). Occasion-Sensitivity: Selected Essays. Oxford University Press.
Zimmerman, T. E. (2020). Fregean Compositionality. In D. Ball and B. Rabern (Eds.), The Science of Meaning: Essays on the Metatheory of Natural Language Semantics (pp. 276–305). Oxford University Press.