13 Semantics

Semantics is the study of meaning, in the sense of linguistic content rather than existential significance. It asks a variety of questions:

These questions and many related ones have direct relevance for the study of religion\s. As scholars of religion, we should be aware of semantics because our work is shaped and constrained by implicit semantic stances, whether we admit it or not. At one level, this is because the phenomena that we study, and the ways we choose to study them, have semantic dimensions. For example, contrasting positions in many debates in the discipline stand or fall as a direct consequence of their semantic presuppositions: e.g. the translatability of sacred scriptures; ‘correct’ methods for studying ritual; the insider/outsider problem; sui generis views of ‘religion’; the status of symbolic or metaphorical language; and the alleged ineffability of mystical experience. At another level, semantics is relevant for all scholarly work of any sort, given that it is shaped and constrained by implicit semantic stances, whether we recognize it or not. All objects and results of study—more generally all meaningful phenomena and all acts of communication—mean what they mean in certain ways, and asking well-informed questions about this has significant practical implications for how we do scholarship. A stumbling block that prevents many people from taking philosophical questions seriously—e.g. epistemological debates about ‘truth’ and semantic debates about ‘meaning’—is that they are so essential and so basic to almost every aspect of our personal and professional lives that we necessarily act on the basis of implicit assumptions that, as a result, are hard to reflect on formally and hard to let go of when proven untenable. In terms of epistemological and semantic views, to offer an analogy that is more appropriate than it might seem at first, most people think that the earth is flat and the sun revolves around it.

A baseline view of semantics sees it as the study of how language is related to the extra-linguistic world, i.e. the study of the relation between our abstract symbol systems and that which they were intended to represent. This would have intrinsic importance to scholars of religion for a number of reasons. Most basically, the religious ‘world’ postulated by the adherent differs from the empirical one that is accessed by both scholars and adherents. The ‘insider’ typically considers the religious ‘world’ to be just as real as (and often coextensive with) the empirical one, but the scholarly ‘outsider’ typically doesn’t. This talk of different ‘worlds’ is meant to represent a commonly held attitude, which this chapter problematizes. This common attitude largely assumes a representationalist semantics: i.e. it is assumed that talk about these ‘worlds’ is fixed by word–world relations, from which it follows that the different ways that insiders and outsiders talk about these ‘worlds’ indicate that they are indeed different. Our point is that interpretationalism—a different semantic view clarified further in the following—doesn’t lead to that conclusion: if ‘meaning’ means something different, the differences between insiders and outsiders are cast in a new light.

In other words, a broader attention to semantics forces us to pay attention to the intentions of speakers and actors in more nuanced ways. Some scholars hold that ‘world’ is a social construction whose basic building materials are the mythological narratives and ritual performances of social groups. A fuller understanding of religions would include what those narratives and performances are supposed to mean. Even to realist scholars—e.g. many theologians who, like adherents, typically regard the religious ‘world’ as objectively real—and to ‘critical’ theorists—who are quite happy to regard it as illusory or even delusionary—understanding and explaining religious phenomena would be greatly advanced by exploring what it is that adherents intend to be representing in their narratives and performances. Empirically grounded approaches that ignore issues of meaning can provide evidence only of the structure and function of religious practices, leaving out what adherents typically take as the most significant or important aspects of their religious lives. Readers might respond at this point that the study of religion\s almost always looks at or takes into account what religious practitioners mean. That is our point precisely—though formal analyses that leave aside issues of meaning in the sense we are discussing here are prominent, e.g. functionalism and syntactical analyses of ritual—and that is why it is important to take semantics seriously. Our point is that methodology, theory, and meta-theory in the study of religion\s could often proceed more effectively and defensibly if pursued in the light of current ‘best practices’ in philosophical thinking about the nature of meaning.

Semantics (the study of linguistic meaning) has traditionally been understood in contrast to syntactics (emphasizing grammar or the rules for combining symbols) and pragmatics (emphasizing how the language is used). In religion, narratives and rituals typically have recognizable syntactic patterns as well as more or less clear pragmatic forces (e.g. functionalist views of religion), but limiting the study of religion to either or both of these impoverishes it. In addition, the distinction between semantics, syntactics, and pragmatics is complicated by advances in the philosophy of language. For one thing, many contemporary philosophers of language challenge whether any significant distinction can be drawn between syntactics, pragmatics, and semantics (as per many of the essays in Ezcurdia/Stainton 2013). On the one hand, the ‘formal’ semanticists regard meaning as a function of syntax (as, for example, the intuitive meaning differences between “Muslims worship only Allah” and “Only Muslims worship Allah”), but pure formalism is incompatible with the fact that communication often succeeds despite improper syntactic construction; on the other hand, ‘use’ theorists regard meaning as determined entirely by how the language is pragmatically used (Davidson 1984), but pure ‘use’ approaches are unable to explain how a never-before-used sentence can be immediately understood (Cappelen/LePore 2006). In a nutshell, language needs compositional syntactic resources, but meaning cannot swing entirely free of its capacity to be understood. Meaning (semantics), structure (syntactics), and use (pragmatics) come as a package.

More challenging to semantics is a split between philosophers of language over whether the role of language is to represent an extra-linguistic world. In other words, many philosophers of language reject a representationalist semantics which sees the meaning of linguistic expressions identified with that which they supposedly represent. For some representationalists, individual words are taken to be signs or symbols, in the sense that they refer to or pick out purported elements of the world. As a variant, some have taken individual words to be signs of inner mental states (‘affections of the soul’ for Aristotle or ‘internal conceptions’ for Locke) but generally have still regarded those mental states as having a representational content. On this view sentences are descriptions of ‘states-of-affairs’ built up out of the objects and properties that its individual words pick out, either directly or as mediated through the mind. The meaning of a sentence is identified with that description: reference determines meaning, and truth—correspondence between these descriptions and the world—becomes a central semantic concept.

Philosopher J. L. Austin’s How to Do Things with Words (1962) fundamentally challenged this approach by bringing to the forefront the fact that we do many more things with words than just describe (see Michaels/Sax, “Performance,” this volume). He drew a distinction between linguistic statements and performances, noting two predominant—but not mutually exclusive—types of linguistic performance: locution (the act of making a statement, as in uttering “I do”) and illocution (achieving some purpose by means of making the statement, such as getting married). However, for Austin it was the statement issued in the locutionary act, not its performance, which carried semantic content, and he continued to think of that content in representational terms (1962, 100).

Non- or anti-representationalists take a much more radical approach, arguing that (i) not all locutionary acts issue statements whose semantic content is given by what they purport to represent, yet (ii) they may still be linguistically meaningful (and so have semantic content). Religions provide powerful examples of (i). The logical positivists’ critique that religious language is meaningless—rejecting (ii)—rests on an assumption of representationalism (e.g. Ayer 1936; Klemke 1960), and so can be easily avoided by taking an alternative semantic stance. The most influential form of anti-representationalism, dubbed ‘interpretationalism’ (see Schilbrack 2014), attributes meaningfulness to anything that can be understood by a neutral ‘interpreter.’

This move has three important elements. First, semantic content would not be limited to linguistic usage that is purely descriptive or truth-evaluable. Types of religious language—prayer, exhortation, incantation, metaphor, etc.—can be analyzed as meaningful even if they lack obvious truth-conditions. For interpretationalists, semantic content shifts from what an expression purports to describe to how the interpreter understands what the linguistic performer is doing. For interpretationalists like Davidson (1984), this involves attributing a range of propositional attitudes to the linguistic performer (e.g. beliefs and desires), that are expressible in the interpreter’s own scholarly language and that would maximally explain why the former is acting the way they act in that context.

Second, meaningfulness need not be limited to ‘language’ narrowly understood in terms of familiar patterns of phonemes and morphemes (e.g. identifiable as such things as English, Cantonese, or Sanskrit). Social-scientifically oriented scholars point out that we can understand—even explain—human behavior generally, and so the number and type of meaning-bearers expand: e.g. including ritual, whether accompanied or unaccompanied by words. Under an interpretationalist semantics, ritualistic behavior, at least that which is intentional, is meaningful. Interpretationalists think of language and linguistic activity in very broad terms, not limited to the familiar morphemes and phonemes of discursive language. This provides a significant philosophical foundation for the shift from belief- to practice-centric conceptions of religion (e.g. Vásquez 2011).

Finally, the shift from representational content (of individual syntactic strings of symbols) to interpretation (of contextually placed and spatio-temporally extended behavior) allows for a holistic rather than atomistic approach to meaning. For interpretationalists like Davidson, linguistic acts, conceived broadly as any instance of intentional behavior, can only be arbitrarily circumscribed: their semantic content is given by a considerable and open-ended range of propositional attitudes attributed by the interpreter to the performer. Such performances always take place in wider contexts, and the more complete the scholar’s awareness of the contextual placement of a linguistic act, the fuller will be her grasp of its semantic content. This implies the inseparability of religion from such things as economy, law, media, nature, medicine, politics, science, sports, and tourism, thus justifying the range of entries in the “Environments” section of this volume.

As an approach to studying religion, the basic framework of any semantic theory is the same: empirically grounded knowledge of the medium (either bits of text for formalists and representationalists or bits of behavior—including text—for interpretationalists), coupled with a philosophically informed theory of what fixes meaning (internal logical structure for formalists, representation of extra-linguistic reality for representationalists, or maximal rationality of overall behavior for interpretationalists), will yield an understanding of the meaningfulness of religious phenomena.

Semantics has had a significant impact on at least two areas directly related to the study of religion\s, namely the philosophy of religion and cognitive science. The typical entities (gods, spirits, fairies, etc.) and attributes (transcendence, omniscience, atemporalness, etc.) of religion parallel the unobservable entities (e.g. electrons) and attributes (e.g. gravity) of science, in that they are postulated rather than directly perceived. A ‘problem of reference’ in the philosophy of science emerges as to what, exactly, terms such as ‘electron’ or ‘gravity’ refer to: are they objectively real objects/attributes that fall outside of our direct observational power (as per scientific realism), or only things that we can directly observe, such as trails in cloud chambers or the acceleration of dropped objects (as per scientific instrumentalism)? A similar, though much older, ‘problem of reference’ emerges in the philosophy of religion, including Aquinas’s analogical approach and Maimonides’s negative theology, as well as, more recently, Ayer’s (1936) denial of meaningfulness, and Alston’s (1988) thesis of univocality. Mysticism provides another nice example in that mystical experiences are often said to be purely private and hence literally indescribable (see Katz 1978). Philosophers of religion have asked whether such experiences could then even be genuine, or if so how it could be verified that two mystics undergo the same experience. These theologically-oriented problems within the philosophy of religion re-emerge as methodological problems in the academic study of religion\s: e.g. on what basis can the ‘outsider’ scholar determine the referents of the language of ‘insider’ adherents (see the following section)?

The cognitive science (CS) ‘revolution’ has stormed the humanities and social sciences, with scholars of religion making important contributions (see Geertz, “Cognitive Science,” this volume). At its core this approach analyzes religious phenomena in terms of cognitive structures housed in the evolutionarily developed human brain. Emerging from an original partnership between psychology, computer science, and philosophy, CS largely views those structures in representational terms and hence as semantically charged. The biggest philosophical contribution to the partnership has been investigation into how cognitive mental states have semantic content. Proponents of ‘embodied,’ ‘embedded,’ and ‘distributed’ cognition argue instead that it is only bodies embedded in social relations that are capable of cognition (e.g. Jensen 2010). This important development has a semantic foundation: the ‘standard’ model tends to assume a representationalist semantics in which the semantic contents of cognitive states are fixed only in their relation to the mind-independent objective world, whereas the newer challenging models think of them in more holistic and socially contextual ways. The standard model makes it difficult to see how the cognitive states are to be accessed by social scientists, or how their content may be sensitive to the contingencies of cultural or historical context. The newer models, on the other hand, better harmonize the obvious advances of CS with the more traditional approaches to studying religion (Gardiner/Engler 2015).

Attempts to define ‘religion’ provide another example of semantics’ centrality to the study of religion\s (see Stausberg/Gardiner, “Definition,” this volume). Not only do differing conceptions of the nature of definition themselves rest on more basic assumptions about meaning, but some scholars have attempted to explicitly define religion as a certain type of semantic structure, the most influential being that of Clifford Geertz: “(1) a system of symbols which acts to (2) establish powerful, pervasive, and long-lasting moods and motivations in men by (3) formulating conceptions of a general order of existence and (4) clothing these conceptions with such an aura of factuality that (5) the mood and motivations seem uniquely realistic” (1973, 90; original emphasis). Notice the emphasis of representationalism in elements (3) through (5), though Geertz does include non-representational factors in element (2) in his definition.

Despite the importance of such issues, semantics is off the radar for most scholars of religion. But there are notable exceptions, scholars fully aware of and explicit about their underlying semantic assumptions: e.g. Lawson/MacCauley (1990) on ritual structures and content along with Stout (1988), Frankenberry (2014), and Davis (2012) on religious practice; Penner (1995) on the role of the ‘truth’ question for scholars of religion; Proudfoot (1985) on religious experience; Godlove (1989; 2014) on religious concepts; and Schilbrack (2014) and Jensen (2003; 2004; 2014) on the subject matter and methodologies of religious studies in general.

More explicit awareness and discussion of semantics in the study of religion\s would move debates on a number of issues to a more fruitful level. For example, if we adopt semantic interpretationalism, then we need not think of religious language as attempting to represent. Rather, on most common accounts, it may be taken principally to prescribe or emote, and as such would not be truth-evaluable. Interpretationalists warn that, once meaning is broken from syntax, we cannot infer that, say, a sentence expressed in declarative grammatical form is necessarily descriptive in its pragmatic force. Even if religious language is declarative in outward form, this tells us very little about what it means. In more radical ways, this opens up the range of relationships between the descriptive and the prescriptive, particularly in allowing prescription to be seen not as something ‘added’ on top of description, but as something that can be quite independent of it. In other words, interpretationalism allows that religious language may be exclusively affective and non-descriptive. For such language, the question of its truth or falsehood need not arise: the ‘truth’ question of religious language is minimized, if not entirely sidestepped. As a corollary, methodological agnosticism, insofar as this is seen as a bracketing of religious truth claims, is largely unnecessary. Furthermore, on non-representational semantics, reference is a function of meaning, not the other way around; i.e. instead of asking what entity a particular name picks out, interpretationalists ask how or whether a range of sentences using a name have overlapping meanings. What is being talked about, then, is determined by what the talk has in common with other talk—with respect to how they can be understood, not their relationship to what exists. As a result, there is little or no ‘problem of religious reference,’ or at least the problem of determining what religious language refers to is no more or less difficult than in the case of any other linguistic expression.

To give another example, if meaning depends upon the pragmatic contexts of language use then, as those contexts vary, there is no such thing as ‘the’ meaning of any religious text, nor does text stand over against ritual as a special site of meaning: rather, meaning reflects context, and intentional actions come to the fore as crucial aspects of those contexts. As a methodological corollary, scholars should take the basic approach whether studying texts or rituals: interpreting them as forms of intentional behavior.

We conclude with three further examples that draw on a specific type of semantic holism, interpretationalism. The holistic perspective on the nature of meaning is dominant among scholars of religion who explicitly discuss semantics (those mentioned at the end of the previous section). Where semantic atomism locates meaning at the level of words, semantic holism locates it in relations between parts of language, from an indefinite network of linked units to an entire language. For example, Lévi-Strauss’s basic insight—that the meaning of mythical classifications is to be found in binary semantic oppositions—is holistic (see Tremlett, “(Post)structuralism,” this volume). Following Saussure’s insistence that language works through semantic difference, Lévi-Strauss shifted emphasis from words or motifs on their own—as if meaning were something to be found in these—to relations between concepts. Structuralism has a holistic view of language: “the linguistic system is not a system of separate items between which certain extrinsic relations hold, but a system wherein the elements of the system are themselves constituted by the differences and relationships between them” (Malpas 1992, 59–60).

One implication of this holistic view of meaning is that it undermines prominent sui generis views of religion. Religion could be argued (if not necessarily cogently) to be a sui generis phenomenon (‘of its own unique kind’) in several senses. We refer here to the most commonly discussed conception (or bugbear) that religion is characterized by something (often a specific type of experience) that is irreducible to non-religious phenomena. The relevant point here is the implication that the meaning of religion is thus methodologically tied to making sense of this irreducible phenomenon on its own terms, not in the same terms used to make sense of non-religious phenomena: e.g. “the meaning of religion … is that it is the inevitable, though by definition inadequate, interpersonal statement and institutionalization of a prior feeling or faith” (McCutcheon 1997, 14–15; original emphasis, paraphrasing W. C. Smith). Because the set of relations that constitute meaning depends on the particular context of a given utterance or bit of text, there is no thing which can be identified as the meaning of any particular portion of language. Judgments regarding meaning are always tentative and open to revision: the interpreter must always concede that there are other, just as well evidenced, ways of understanding a given meaningful expression or act. Many semantic holists take this further by rejecting the possibility of fundamentally different types of meaning or interpretation. Davidson, one of the most influential of holists, famously rejected the idea that there can be such a thing as metaphorical meaning distinct from literal meaning; and the same critique applies to the possibility of some form of distinctly transcendent or symbolic meaning that supposedly emerges from some special relation between religious language and its ‘sacred’ referents (Davidson 1984; Frankenberry 2002; Penner 2002). In sum, if there is only one type of meaning, then sui generis views of religion (in the above sense) are non-starters.

A second implication of a holistic view of meaning is that religious language cannot be sharply distinguished between other types of language, e.g. scientific, moral, aesthetic, or political. The meaning of religious terms and concepts is cashed out in terms of their relation to others, and these others are those that happen to appear along with the religious ones:

Semantic holism explicates meaning by reference to positions or roles in a vast web of interconnected points whose portions do not admit of discrete dissection. Understanding the claim that Yahweh called to Moses from a burning bush blends items from traditionally different subject matters, e.g., Jewish scriptural tradition, botany, and pyrogenics. The meaning of each those nodes is given by relation to still others, and so on.

(Engler/Gardiner 2010, 289)

Whether one accepts semantic holism or not, there is an important general question here that can only be answered by paying attention to semantics: “The question seems to be whether there really is any such specific entity as ‘religious language’ and/or whether the semantics of religious systems are just ‘plain’ semantics of an order similar to other specialized terminological systems, those of, say, politics, sports or economics” (Jensen 2004, 220).

A third implication is that there can be no strong insider/outsider problem (Gardiner/Engler 2012). A holistic view of meaning leaves scholars with a weak or relative problem—the methodological challenge of studying groups that have subject matters and vocabularies that are relatively inaccessible to outsiders—but not an absolute one—the claim that religious insiders have unique and monopolistic access to religious knowledge. The latter is associated with sui generis views of religion (see earlier discussion). If the meaning of expressions is a function of their relations to the meaning of others (as in semantic holism), then the meaning of insider language is arrived at in part through easier-to-interpret non-religious language, which is shared by outsiders. In other words, because religious language is not some distinct, walled-off sphere of meaning, because it necessarily overlaps into other types of language that are shared by non-insiders, scholars may have to work hard in their attempt to understand insider discourse (a relative challenge), but that is a practical hurdle not an inescapable dead-end (not an absolute barrier to research and interpretation).

The study of religion\s has been little informed by semantic approaches. The ‘linguistic turn’ in early twentieth-century Western philosophy—in both the ‘analytic’ branch (e.g. Frege, Russell, Wittgenstein, Austin, Carnap, Quine, Davidson, Rorty) and the ‘contintental’ one (e.g. Saussure, Heidegger, Gadamer, Derrida, Ricouer, Foucault, Habermas)—explored how philosophical reflection on meaning constrains approaches to philosophy’s traditional concerns, especially epistemology, metaphysics, and ethics. A predominant debate in ethics, for example, concerns whether ethical language points us to objective moral truths or only to expressions of subjective inner sentiments; when a vegetarian declares “Eating meat is wrong,” is she best understood as stating a moral truth (or falsehood) or only expressing, in verbal terms, her love of animals, or something else entirely? A significant debate in the philosophy of science is over what specialized scientific vocabulary actually means: can scientists really refer to unobservable entities, like the Higgs boson, and if not can they really be thought to exist? Do biological species-terms pick out natural kinds, or are they rather to be understood as pragmatically fruitful shorthands for ultimately amorphous groupings? Our very conceptions of ethics and science are tied to our conceptions of what we are doing when we use language ethically and scientifically, and those conceptions are inseparable from an understanding of how those uses might be meaningful. Analogously, the scholar of religion cannot separate a conception of religion from a conception of what adherents are doing when they use religious language (either narrowly as per the formalists/representationalists or broadly as per the interpretationalists). Contrary to Frits Staal’s contention that much religious ritual is ‘meaningless’ (1979), adherents clearly intend their religious language to mean something, and the nature of that meaning should guide the scholar’s theoretical reflections and methodological practices.

The importance of semantics to the study of religion is manifold. On an object level, approaching religious phenomena as being either analogous to or describable in linguistic terms has proven very fruitful for scholars of religion. This approach has opened up many interdisciplinary doors and methodologies (e.g. linguistics, generative grammars, linguistic anthropology, cognitive science, semiotics, discourse analysis, etc.). Ignoring the semantic side—that linguistic expressions mean something—impoverishes the field. At a theoretical level, it is important to understand (i) that there are a range of divergent but philosophically respectable semantic positions, and (ii) that the choices that one makes between them, either implicitly or explicitly, have implications with respect to general theoretical frameworks that should guide one’s research. On a meta-theoretical level, the admissibility and plausibility of this-or-that bit of research will stand or fall with the admissibility and plausibility of the semantic assumptions it makes.

While scholars of religion needn’t engage in direct philosophical defense or critique of semantic frameworks, they should be aware of at least the main contours of different positions and to acquire a minimal working knowledge of what’s happening in the philosophy of language. Every scholar of religion—regardless of their academic home, cultural/historical focus, or theoretical preference—makes semantic assumptions that constrain which ways of thinking are in and which are out of bounds.

 
• Atomism/holism

  a basic divide in semantic theories; atomists hold expressions have their meanings intrinsically, whereas holists hold that those meanings depend on their relations to the meanings of other expressions.

 
• Contextualism/formalism

  a basic divide in semantic theories; contextualists hold that the meaning of linguistic expressions is largely determined by the pragmatic contexts of its usage, whereas formalists regard them as largely determined by their internal syntactic structure.

 
• Interpretationalism

  an influential holistic and non-representationalist semantic theory that identifies meaning with the range of propositional attitudes attributable to some subject by a neutral interpreter, in order to maximally explain the subject’s overall intentional behavior.

 
• Meaning

  semantic content, information conveyed (as opposed to personal or subjective importance).

 
• Monism/pluralism

  a basic divide in the theory of meaning; monists hold that all meaningful expressions are meaningful in virtue of sharing the same underlying property, whereas pluralists hold that there are several and distinct types of meaning.

 
• Reference

  the relation between a name and its bearer (= referent); theories of reference are theories of what fixes that relation, including descriptive theories (reference is fixed by whatever satisfies an implicit description), ostension (reference is fixed by pointing at the referent), and causal theories (in which the referent stands in a direct causal relation to its name, as in a case of baptism).

 
• Representationalism/non-representationalism

  a basic divide in conceptions of the basic function of language; representationalists regard language inherently as a system of representing or describing non-linguistic ‘states-of-affairs,’ whereas non-representationalists deny that this is the central or even a significant function of language.

 
• Semantics

  the study of meaning, in the sense of linguistic content rather than existential significance.