1. Introduction
The significance of studying religious language lies in its effort to determine how to appropriately characterize the divine, an entity regarded as the transcendent source of all creation. Traditionally, numerous theories have engaged in the debate over this issue, specifically focusing on whether the literal or metaphorical usage of words and meanings within religious language is sufficiently effective. In this paper, I address a gap in the existing literature; this is a thorough study that reviews the definitions of literality and metaphoricity in religious language and, subsequently, proposes a new framework to advance the field. To achieve this, I will review both traditional and contemporary theories, demonstrating how the insights derived from this review can help us develop the subject a step further.
The problem of religious language arises from a complex set of assertions central to the main philosophical theologies within religious traditions that believe in a transcendent God, such as Judaism, Christianity, and Islam (see
Burrell 1993). According to these traditions, a transcendent divine entity, by definition, lies beyond anything in the world and, thus, beyond human understanding and language. Yet, this entity is also the creator of the world and, as such, remains connected to us. For believers, the transcendent source of the world is of the utmost importance, prompting a desire to understand and characterize the divine in ways that can be communicated and shared with others. However, any language we use to characterize this entity is inevitably limited and cannot fully capture its nature. This challenge brings us to the question of how, and to what extent, language can genuinely refer to it. Among the various expressions, the perfection terms stand out as some of the most viable options for referring to the divine, as human perfections are understood to be reflections of divine perfections. For example, consider the term “good”. We might ask what does goodness mean in humans, and what does it mean in God? In what ways is God’s goodness similar to or different from our own? Is God good in the same way that a friend is good—by caring for us, being kind, or exhibiting physical attractiveness and moral behavior? A simple response might be that God is good, but not exactly in the same way as a human friend, because God lacks the corporeal limitations that human friends have. The phrase “not exactly” is intended to remove any implication of limitation or restriction from divine goodness, suggesting instead that God performs good deeds in a better and more fulfilled way than others. A more technical answer suggests that divine goodness is reflected in the act of creation, where God grants existence, the foundation upon which all human perfections are built. This raises deeper questions about the nature of existence itself and how it relates to the existence that is ascribed to God. These kinds of answers open the door to more intricate formulations, which require a thorough understanding of the language itself—a topic that will be briefly categorized in the following paragraph.
One traditional view, which finds support in modern thought, particularly among those influenced by Wittgenstein’s early work (
Clack and Clack [1998] 2019, pp. 110–11),
1 advocates the idea that language concerning such a being is inherently nonsensical, leading to a form of quietism.
2 It is noteworthy that this
Tractarian nonsensicalist picture proposes a “thin” view of religious language and raises the question of why one might engage in God-talk at all (
Andrejč 2016, pp. 33–34).
3Another perspective asserts that we can characterize the divine using the same language and contexts as ordinary discourse, endorsing a literal interpretation and contending that metaphorical statements are deviations or illegitimate forms of literal language (see the substitution and comparison theories explained in
Section 2). A different proposal asserts that, due to the radical distinction between the divine and the human, religious language is strictly metaphorical and figurative, devoid of any literal element. Some scholars, such as
McFague (
1983) and
Kenny (
2004, pp. 40–41), arguably support this perspective. A fourth proposal extends beyond, arguing that both the metaphoricity and literality of religious language are to be played. This perspective has various interpretations in the literature. Theologians like
Burrell (
1973,
1979) emphasize the metaphorical aspect, while still acknowledging the literality of
via analogia. In a similar effort, drawing on altered focus in linguistics, such as those by
Höfler and Smith (
2009) and
Gibbs and Tendahl (
2011), some argue that distinguishing between literal and metaphorical language is not only undesirable but also impracticable. Among particular recent scholars, there remains a distinction, with some defending the genuineness of the literal side, such as
Allott and Textor (
2022), and others supporting the metaphorical side, like
Hesse (
2023).
In this article, I will examine the second and third proposals, followed by a relevant scholarly discussion on the fourth proposal. I will not engage with the first proposal, as it eludes capture by any conceptual scheme. Through the methods of conceptual analysis, grammatical inquiry, and theological reasoning, I aim to demonstrate that what I refer to as “analogical dialectics” offers a more reasonable interpretation of the distinction in question. I propose this notion because I believe it answers the question of why existing accounts of religious language are so ambivalent, moving back and forth between the notions of literality and metaphoricity. This approach will also serve to elucidate, address, and sidestep the problematic issues associated with the second and third proposals.
2. Conventional Interpretations
Contemporary scholarly debate on metaphor often positions itself in response to conventional views, rejecting traditional interpretations and revising early modern perspectives. Traditional accounts, such as substitution and comparison theories, treat metaphor as a deviation from literal language, asserting that metaphors can be translated into literal expressions without losing their meaning. Substitution theory, attributed to Aristotle and Quintilian, regards metaphors as rhetorical devices that can be replaced by literal terms. Comparison theories, including the most recent versions offered by
Robert Fogelin (
2011) and Amos Tversky, suggest that metaphors only highlight resemblances between entities; however, they are criticized for assuming pre-existing similarities and overlooking the role of metaphors in generating new meanings (
Hills 2022).
The situation shifted somewhat in the 20th century, as numerous theories on the literal–metaphorical nexus emerged from an interdisciplinary context, drawing from fields such as linguistics, psychology, philosophy of language, linguistic philosophy, and theology. These theories deliberately distanced themselves from traditional reductionist tendencies and laid the groundwork for contemporary perspectives by incorporating post-Kantian and Wittgensteinian insights that questioned the relationship between language, mind, reality, and religion. Given that these theories offer valuable insights for the development of my proposal in this article, I will discuss the most relevant ones.
A key approach that illustrates this shift is interactionist theory, as exemplified in the works of
I.A. Richards (
1979),
Max Black (
1954), and
Harold Skulsky (
1986,
2012), which highlight the role of a semantic twist in metaphors. It understands metaphor as a filter through which we perceive objects, revealing the structural correspondences between different concepts, continuously generating new insights, and altering our understanding of both metaphorical and literal contents. This theory posits that metaphors emerge from the interplay or “interanimation” of words and their meanings as they interact within the context of specific utterances, creating a “tension” between different layers of meaning. When we take an assertion as a metaphor, we assign it a new and distinctively metaphorical meaning, with the meanings of focal words undergoing a metaphorical twist. The latter is produced by the more unyieldingly literal meanings of the expressions that serve as the metaphor’s frame (see
Hills 2022;
Hector 2011).
Hector (
2011, p. 117) discussed two main types of criticisms, with the second being the one he finds more compelling. First, this view generates dual meanings: literal and metaphorical. The literal meaning often appears absurd, prompting a search for a metaphorical interpretation, suggesting that metaphorical meaning arises from interpreting this literal absurdity. Second, critics claim that interaction theorists confuse subjective responses with objective meanings, similar to restating a function without explaining it. Thus, it is stated that the interactionist view is ultimately bound by a fixed, literal core, treating metaphorical meaning as secondary, deviant, or decorative.
While being criticized, the interactionist account of metaphor contains key ideas that remain valuable for my discussion. Among its merits is the concept of tension, initially proposed by Black and further developed by other interactionists like Paul Ricœur. In Black’s framework, tension arises from the interaction between the source and target domains, creating a dynamic interplay that generates new meanings.
Ricœur (
1978), though, locates the conflict that gives rise to metaphor in the incongruity found in a literal interpretation of nonsensical words. He argues that a metaphor exists “in and through an interpretation”, where the literal sense connected to a literal reference produces what he calls an “absurdity” (
Ricœur 1976, p. 50). This absurdity creates tension between a literal and metaphorical interpretation. Ricœur explains, “The metaphorical interpretation presupposes a literal interpretation which self-destructs in a significant contradiction”. This process of self-destruction or transformation “imposes a sort of twist on the words”, leading to an extension of meaning that allows us to make sense where a literal interpretation would be “literally nonsensical”.
4 The importance lies in the tension itself, as it offers a vivid contrast between two domains, merging the most abstract concepts with the most concrete realities, thus enriching our cognitive and perceptual frameworks (see
Indurkhya 1992, pp. 74–75).
Another expansion in the discussion comes from H.P. Grice, with further development by John Searle, who attempt to respond to the rationale behind such a contrast by proposing that metaphors involve saying one thing while meaning/implying another. This approach relies on pragmatic conversational implicature and the cooperative principle. Grice’s theory suggests that when conversational maxims are intentionally violated, listeners infer additional meanings, leading to metaphorical interpretations. It suggests that metaphors possess both propositionality and truth-conditionality (
Hills 2022). Building upon this, Searle argues that there is a systematic process involved in how metaphorical meaning is derived when speakers intentionally violate conversational norms. He emphasizes that meanings are contextually fine-tuned, with modulation occurring within the norms of correct usage. This challenges the idea that nonliteral uses are fundamentally different from literal ones, suggesting that meaning modulation is common across all language use (
Searle 1980,
2012). Searle’s notion of fluid boundaries between literal and nonliteral meanings was later adopted and further developed in relevance theory, which I will discuss in subsequent sections.
Donald Davidson, however, rejects the idea that metaphors convey additional propositional content beyond their literal meaning. Instead, he argues, metaphors function by inducing an effect, prompting listeners to view the primary subject in a new light without attributing specific metaphorical meanings to the words used. Davidson’s theory, which underlines the role of “brute force”
5 in metaphors, is influenced by Wittgenstein’s perspective on contextual understanding, the importance of pictures, and the distinction between “seeing as” and “seeing that”. Seeing as involves a perceptual shift that is not fully captured by propositional statements; whereas, seeing that can be expressed in plain words. For example, the “duck–rabbit” illustration shows that recognizing different aspects—seeing as—is distinct from merely acknowledging a fact—seeing that (
Davidson 1978, p. 47).
6 Davidson posits that metaphors do not have a separate metaphorical meaning that can be paraphrased or translated into literal language. Instead, the power of a metaphor lies in its ability to influence how we perceive and think about the subject. For example, the metaphor “Juliet is the sun” does not imply a direct comparison or substitution but rather encourages the listener to consider Juliet in terms of the qualities associated with the sun. Davidson also elaborates on Morgan’s concept of the “framing effect” (see
Moran 1989, p. 98), elements that prompts listeners to see one thing in terms of another, in which a novel perspective is created through the juxtaposition of primary and secondary subjects. Davidson simulates this to the process of understanding “jokes”, which require wit and a sense of humor, where the literal meanings and syntactic roles of words remain unchanged. This approach proposes that the strength of metaphors lies in their ability to foster comparisons and analogies, unveiling new insights to a listener through “special nonstandard communicative use” (see
Hills 2022) without relying on hidden messages and implicature.
The theories discussed so far in this section suggest that, while literal and metaphorical elements are both genuinely present in our language, they are not entirely compatible, and this partial incompatibility leads to the emergence of new meanings from their interactions. Certain theological schools of 20th century have incorporated both elements into their discussions, notably those integrating post-Kantian and Wittgensteinian interpretations into Thomas Aquinas’ account of religious language. Among these interpretations of Thomism, the literal versus metaphorical debate was central to a broader discourse on apophatic–cataphatic approaches to the
via analogia. The apophatic process entails the negation of any characterizations of God that are corporeal or imply limitation, while the cataphatic approach, conversely, seeks to establish an affirmative connection between God and creatures. Through the cataphatic method, one approaches more literal characterizations of the divine; whereas, the apophatic method inclines toward metaphorical expressions, as it endeavors to transcend the constraints of literal language and reach a more proper understanding of the transcendent divine. Some Thomists still adhere to a form of reductionism. For example, some view religious language as highly metaphorical, such as
Lonergan (
1963) and
Rahner (
1987), known as transcendental Thomists, and
Kenny (
2004) from the analytical Thomism camp. On the other hand, most analytical Thomists, including
Geach (
1950), advocate for the literality of religious language (
Kenny 1959). However, certain theologians depart from these reductionist tendencies and attempt to develop a middle way. These include
Burrell (
1973,
1979), a grammatical Thomist, as well as David Tracy and Erich Przywara (
Tracy 1987;
Przywara 2014) from the so-called hermeneutical camp. In this article, the perspectives of the latter group will feature, as they are particularly relevant to the discussion.
To begin with, Burrell’s extensive discussion of metaphor differentiates between metaphorical and literal language by stating that metaphors involve figurative or symbolic comparisons that extend understanding by highlighting similarities in new or obscure concepts; whereas, literal language aims for precision and unambiguous description, particularly in scientific discourse. He discourages the pursuit of literal precision, arguing that metaphorical constructions are indispensable even in fields demanding exactness. At the core of analogical language, he suggests, lies a metaphorical dimension of a vague kind, allowing analogical predictions to respect the distinction between God and creation. Burrell argues for the irreducibility of God-talk to purely literal language, asserting that theological discourse cannot be confined to precise, unambiguous terms. Yet, he also acknowledges that religious language must also be literal. Burrell’s work, however, plays with the different meanings of literality and metaphoricity. When he confirms the apophatic aspect of analogical predication, he refers to religious language as strictly metaphorical, terming it paradoxical language (
Burrell 1973, p. 266). This language is not literal in any conventional sense. When Burrell pronounces that analogical religious language with a metaphorical core is literal, he employs the terms “literal” and “metaphorical” in a nonstrict sense. Yet, another interpretation of “literal” is said to be that which resists metaphorical reading.
7 Stephen Mulhall (
2016, p. 63), in his interpretation of Burrell’s account of analogy, suggests that religious language is a riddle, implying that both literal and metaphorical elements are mysteriously intertwined.
Two other theologians warrant reference for their explicit articulation of the concepts I wish to emphasize: David Tracy (b. 1939) and Erich Przywara (1889–1972). Although Przywara should have been discussed earlier in this paper based on chronological order, I address him here due to his particular relevance to the current topic.
Tracy is a Catholic theologian associated with theological hermeneutics and fundamental theology. Influenced by Mircea Eliade, Hans-Georg Gadamer, Bernard Lonergan, Karl Rahner, Paul Ricœur, and Paul Tillich, Tracy has left a mark on contemporary theological discourse, especially religious pluralism and comparative theology (
Tracy 1987). He advocates for analogical language as a vital tool for discussing God, emphasizing that understanding the divine benefits greatly from both transcendental Thomism and process thought, which share a common departure point in human experience. His hermeneutical approach reflects his shift from phenomenology to a more historically and culturally situated understanding of religious experience, aligning with Gadamer’s insights into the nature of understanding.
Tracy examines the concept of “negative dialectics”, drawing from Protestant neoorthodox theology to expose and challenge theological illusions and pretensions through a logic of contradiction. For Tracy, negative dialectics involves a critical examination of theological concepts, revealing the limitations and contradictions inherent in the human understanding of the divine. He emphasizes the importance of incorporating the insights of negative dialectics into analogical religious language to avoid theological sterility. By this approach, negative dialectics serves to ensure that theology remains dynamic and self-critical, preventing it from falling into simplistic or dogmatic assertions. Tracy argues that the failure to integrate negative dialectics can result in theological language becoming univocal or dissipating into pure equivocality (
Tracy 1983). By re-examining fundamental theological concepts, such as God’s relationship with the world, theologians can remain faithful to both religious and logical principles. While similar discussions appear in the work of other Thomists under the notion of negative theology, the significance of Tracy’s approach lies in his distinctive emphasis on the dialectical aspect of God-talk.
The role of the dialectical element in religious language is emphasized and elaborated in the works of Erich Przywara (1889–1972). Although he was not widely recognized, his thought exhibits notable depth and left an influence, particularly within the schools of phenomenology and Thomism. Przywara’s concept of the “analogia entis” (analogy of being) is central to his philosophical and theological thought, where he presents it as a dynamic and dialectical tension between God’s immanence and transcendence.
8 His work illustrates the characterization of the divine grounded in a theological understanding of the concept of existence, offering a potential answer to the problem of religious language, as exemplified in the introduction of this article. According to Przywara, the analogy of being is not a static relation but a continuous, oscillating movement between the divine and the created order. This “unity-in-tension” posits that while God is wholly other and beyond all creation (transcendence), He is simultaneously present and active within it (immanence). Przywara articulates this relationship as a “suspended middle”, where the divine essence and its manifestations in the world are held in a state of perpetual balance, neither fully resolving into a singular unity nor collapsing into pure duality (
Przywara 2014, p. 211). This tension, for Przywara, reflects the intrinsic dynamism of reality and underscores the limits of human understanding, which must continually navigate the paradoxes inherent in the relationship between the finite and the infinite (
Przywara 2014, p. 214). In
Section 5, I will highlight the importance of Przywara’s dialectical analogy for explicating the paradoxical relationship between literal and metaphorical interpretations.
To conclude this section, the developmental trajectory of scholarship on language reveals that the evolution of language and the creation of new meanings are deeply rooted in the interaction between literal and metaphorical meanings. Furthermore, while there are occasions in which this tensive interaction may diminish in the context of ordinary language use, it remains particularly dynamic in the realm of religious language. In the next section, I will draw on the most recent theories to further explore and clarify the characteristics of this interactive relationship, shedding light on how these dynamics shape our understanding of religious language.
3. Renewed Interest in Metaphor Theory
Lakoff and Johnson’s introduction of conceptual metaphor theory (CMT) in their seminal book,
Metaphors We Live By (
Lakoff and Johnson 1980), marked a significant paradigm shift in contemporary scholarship’s understanding of metaphor theory. In light of this new understanding, the notion of “literality” warrants a re-evaluation, which I will start with in the discussion that follows.
Lakoff (
1986, p. 3) identifies four theory-oriented accounts of literality. It is crucial, he suggests, to avoid conflating them and to maintain their distinctions, as a particular expression may be deemed literal under some accounts but nonliteral under others, as illustrated by the definition of electricity. The first account is referred to as conventional language literality, which pertains to the conventional use of language, as opposed to “poetic language, exaggeration, approximation, embellishment, excessive politeness, indirectness, etc.” (
Lakoff 1986, p. 2). For example, describing electricity in terms of fluid flow, such as in the sentence “a resistor will dam the flow of electricity”, is literal, as it fits within conventional language. The second type is called subject matter literality, which pertains to language used explicitly to discuss particular subjects with relevant terms and phrases. For example, the crowd metaphor for electricity, as in “a resistor is a narrow gate that will only let a certain number of electrons get through at once”, illustrates subject matter literality. The third type is nonmetaphorical literality, which involves a direct and meaningful application of language without needing interpretation through another context. Therefore, describing electricity in terms of fluid flow or crowd cannot not be seen as an example of nonmetaphorical literality. Finally, the fourth type introduces truth-conditional literality, which reflects reality, referencing actual entities or proclamations that can be decisively deemed true or false. Since the example of electricity is not objectively true in the literal sense, as we cannot directly observe electricity and its workings, it does not fit truth-conditional literality.
9 Lakoff critiques traditional theories for confusing some of these meanings, which leads to several misunderstandings. This conflation often results in conventional literality being mistaken for nonmetaphorical literality, which implies that conventional language is direct and devoid of metaphor. Additionally, conventional language literality is frequently seen as equivalent to truth-conditional literality, suggesting that ordinary language use can accurately reflect objective reality. Moreover, subject matter literality is often presumed to be identical to truth-conditional literality, implying a singular, objectively correct interpretation for each subject matter (
Lakoff 1986). According to this account, the nonmetaphorical literality signifies a strict “directness” in linguistic application, thus circumventing the necessity for metaphorical use. However, conventional language literality, subject matter literality, and truth-conditional literality can coexist with metaphorical meaning. Given these considerations, Lakoff proposes that the term “literal” should be primarily associated with the third category—nonmetaphorical literality. Nevertheless, Lakoff dismisses the fourth reading, namely truth-conditional literality, arguing that this theory is based on a flawed understanding of the nature of metaphoricity and literality. This understanding of literality is a result of conceptual metaphor theory, a framework copioneered by Lakoff and Mark Johnson (
Lakoff and Johnson 1980). Let me now touch upon CMT and certain other contemporary theories and discuss the way they can further the topic at hand.
A conceptual metaphor, according to the CMT, is originally a cognitive process that generates new images and ideas through “metaphorical mapping”. It is crucial to understand that conceptual metaphors primarily operate unconsciously, shaping our conscious thoughts. Conceptual metaphors, such as LIFE IS A JOURNEY and ARGUMENT IS WAR illustrate the foundational structures from which many other metaphors are derived (See
Lakoff and Johnson 1980, pp. 253–56;
1999, pp. 12–13). While humans’ engagement with linguistic metaphors tends to be more cognitively aware, there remains a strong connection between the two. Metaphorical linguistic expressions act as representations of conceptual metaphors rooted in cognitive processes, where elements from one context, referred to as the “source domain”, are mapped onto another context, known as the “target domain” (
Lakoff and Johnson 1980, pp. 253–56).
To explore the relationship between metaphorical and literal language as defined by CMT, it is necessary to revisit conventional language literality and nonmetaphorical literality in more detail. The treatment of nonmetaphorical literality in cognitive theories is primarily featured in discussions on the role of nonmetaphorical meanings in forming “primary metaphors” (see
Lakoff and Johnson 1980, pp. 253–56;
1999, pp. 12–13). Primary metaphors involve projecting image schematic patterns onto abstract concepts. Linguistic expressions that embody these patterns from our perceptual experiences (e.g., spatial experiences, such as “container”) are considered literal. These expressions are deemed nonmetaphorical, because we understand their meanings independently, without relating them to other contexts (see
Lakoff and Johnson 1980, pp. 253–56;
1999, pp. 12–13). It is noteworthy, however, that not all contextual adaptations lead to the figurative use of language; there are statements that are contextually adapted yet still regarded as literal language.
10 This type of literal meaning constitutes a small portion of our everyday language, while a larger portion is composed of metaphorical extensions (
Lakoff 1986, p. 3). A large part of statements used by language users in ordinary communication are essentially metaphorical but are often perceived as literal expressions. The literality of these statements, referred to as conventional language literality, stems from the ease with which users understand their meanings without further reflection. For example, phrases like “going to”, “being under pressure”, or “feeling really down” are idiomatic expressions that have become so embedded in English that their metaphorical nature is often overlooked, leading us to interpret them literally. Cognitive linguists refer to these as “dead metaphors”, because they have become conventionalized and acquired distinct literal meanings (
Lakoff and Johnson 1980, pp. 211–12). When a metaphorical usage becomes lexicalized, an interpretation that relies on the explicit inclusion of this concept in the content is considered strictly literal (
Sperber and Wilson 1995, chap. 4). Lakoff concludes that “Only literal-3 [nonmetaphorical literality] contrasts with metaphorical” (
Lakoff 1986, p. 5).
Nevertheless, in relation to Lakoff’s discussion, there remains another “contrast” that he initially mentioned but did not revisit. At the outset, he stated that conventional language contrasts with “poetic language, exaggeration, approximation, embellishment, excessive politeness, indirectness, etc.” (
Lakoff 1986, p. 2). Given his recognition of the ubiquity of metaphorical uses within conventional language and his final remarks, it is more appropriate, I propose, to reorder his discussion as follows: There are two types of metaphoricity. One type involves a kind of usage (e.g., poetic or highly figurative use), which contrasts with both conventional language literality and nonmetaphorical literality. Conventional language, our typical mode of communication, includes conventional metaphors that can be contrasted with nonmetaphorically literal assertions. Additionally, several metaphors are regarded as dead or lexicalized and consequently literal, depending on the context in which they are used. Besides, one might question why the same cannot be said for highly figurative language employed by, for instance, a community of poets, where some imaginative statements might be considered dead metaphors, while remaining highly figurative in other contexts.
In light of these observations, it becomes evident that distinguishing between the boundaries of literal and metaphorical meanings is challenging, leading some to argue that cognitive semantics does not aim to differentiate between literal and figurative meanings. However, in elucidating the embodiment or “hardwiring” features of cross-domain mapping within our cognition, a cognitive semantic approach struggle to identify the literality of elements, whether they are strictly nonmetaphorical or lexicalized (See
Harkaway-Krieger 2024, p. 12;
Masson 2014). I suggest that CMT effectively illuminates one side of the coin, while the other side—language in use—remains to be addressed by other theories.
The way language operates in communication, usage, and practice tailors the claim that language is originally rooted in cognitive systems and, therefore, inherently metaphorical, as suggested by CMT. Recent studies (see
Section 4) highlight the difficulty in distinguishing between literal and metaphorical interpretations, a challenge that extends to religious language, where theology’s primary task is to convey insights and arguments about God-talk in a manner that is both meaningful and convincing. In the following section, I will examine how this communicative aspect of language necessitates moving beyond the metaphorical–literal dichotomy.
4. Beyond the Literal/Metaphorical Reductionism
In this section, I argue that, to gain a more nuanced understanding of how the literal-metaphorical process in language arises, develops, and functions in social communication, the first step is to examine what can be learned by integrating relevance theory with CMT, as suggested by recent scholars. I also demonstrate that the literal and metaphorical cores of language cannot be reduced to each other, nor can they be clearly distinguished from one another. Furthermore, I will indicate that this perplexity is even more pronounced in religious language.
The deflationary approach to linguistic representations is supported by relevance theory, which builds upon and refines the Gricean pragmatic framework (
Hills 2022). Dan Sperber and Deirdre Wilson’s early formulation of relevance theory provides a novel framework for understanding how both literal and nonliteral meanings are constructed through the same inferential processes (
Sperber and Wilson 2008). Literal interpretations, they suggest, are not the default mode but rather one among many possible interpretations guided by the principle of relevance. The latter proposes that human cognition is geared towards maximizing relevance, meaning that the interpretation process seeks the most contextually appropriate meaning with the least cognitive effort. Literal and metaphorical meanings are generated by the same mechanism, the narrowing–broadening process. Concept narrowing refers to focusing on specific, contextually relevant aspects of a word, while concept broadening involves extending a word to related concepts. This dynamic process creates ad hoc concepts that are tailormade for the specific context of the utterance. For example, the phrase “I have a temperature” narrows the concept of “temperature” to imply a fever, which is contextually relevant in a discussion about illness. Conversely, when said “Holland is flat”, the concept “flat” is broadened to encompass a relative flatness suitable for cycling, even though it is not perfectly flat (
Sperber and Wilson 2008, p. 91).
Sperber and Wilson further argue that there is a continuum between literal, loose, and metaphorical uses of language, with no strict boundaries separating these categories. These three kinds of interpretations are derived through the same inferential mechanisms, differing only in degree rather than kind. Metaphors, according to their deflationary account, are not exceptional but are a natural extension of the way we use language to communicate effectively. For instance, the metaphor “this surgeon is a butcher” emerges through inferential processes that leverage background knowledge about the typical actions associated with butchers and surgeons. This inferencing evokes the emergent property of incompetence, demonstrating that metaphors arise from meaningful cognitive associations rather than arbitrary connections. Demonstrating that such a continuum operates in semantic extensions suggests that traditional views, which treat metaphors as deviations from literal language, fail to capture the inherent fluidity and interconnectedness of linguistic interpretations (
Sperber and Wilson 2008, pp. 97–98).
Another key aspect of the relevance theoretic approach is rethinking the significant role of context in shaping meaning. Human communication is fundamentally inferential, where the speaker provides evidence of their intended meaning, and the addressee infers this meaning from the context and the evidence provided. This inferential process is guided by the communicative principle of relevance, which holds that every act of communication conveys a presumption of its own optimal relevance. In practical terms, this entails that the communicator’s message is presumed to be worth processing and expected to be more relevant than competing inputs. This principle applies equally to literal and metaphorical language, thereby challenging the notion that metaphors require a special interpretive mechanism. Strictly literal interpretations, which neither narrow nor broaden the lexicalized concept, are arrived at through a single mutual adjustment of explicit content and implicatures. Sperber and Wilson argue that the apparent “fuzziness” and overlap among literal, loose, and metaphorical utterances indicate that there are no genuinely distinct categories from a descriptive, psycholinguistic, or pragmatic perspective. More importantly, all these types of utterances are interpreted using “the same inferential procedure” (
Sperber and Wilson 2008, p. 95). By recognizing that both literal and metaphorical meanings are constructed through the same relevance-driven inferential processes, the deflationary account offers a unified interpretation of how one can understand different linguistic extensions.
Building on the relevance theoretic approach, Höfler and Smith argue that metaphorical language arises when an existing linguistic form is employed to express a meaning (i.e., its metaphorical meaning) that shares similarities, though not identical, with the one conventionally associated with that form (i.e., its literal meaning) (see
Höfler and Smith 2009). They contend that there is no definitive breach between literal and metaphorical usage; rather, they exist on a continuum. The divergence between the literal and metaphorical use of language is a matter of the degree of similarity between the conventional meaning of the linguistic form and the meaning it seeks to convey. While metaphorical language necessitates a degree of semantic shift, it should not entail a substantial divergence that breaks from the original meanings but rather incorporate minor deviations. This is prevalent not only in poetry or religious language but also in everyday communication. Metaphorical language requires overlooking some differences—deemed irrelevant—between two things to emphasize a pertinent similarity for a specific communicative purpose, reflecting continuity with any language application, where aspects of the conventional meanings are overlooked within the specific context. This interpretation applies not only to conventional metaphors but also to new and creative metaphors being formed within language (see
Höfler and Smith 2009).
CMT and relevance theory, as we have seen, present several differing focal points. In the former, conceptual metaphors are prioritized over linguistic representations; whereas, in the latter, particularly in Wilson’s more recent work (
Wilson 2011), linguistic expressions are posited to have priority, suggesting that conceptual metaphors may emerge from verbal metaphors. Despite these differences, the potential for fruitful interaction between the two approaches has prompted some scholars to pursue a unifying framework.
Gibbs and Tendahl (
2011), for example, contend that these theories offer more complementary insights than previously recognized. They argue that conventional cross-domain metaphorical mappings, such as LIFE IS A JOURNEY, may become entrenched due to repeated verbal metaphors, indicating a bidirectional relationship between metaphoric thought and language. This emphasizes the importance of social and contextual factors in our understanding of metaphors in real-time communication, indicating that metaphoric understanding is a product of cognitive processes coupled with communicative interactions. As a result, repeated verbal expressions can reinforce metaphorical mappings, highlighting the intricate connection between metaphorical cognition and linguistic expression (
R. W. Gibbs and Tendahl 2011, p. 606).
There are still some voices challenging this unificatory account. For example, Nicholas Allott and Mark Textor argue for maintaining a clear distinction between literal and metaphorical language, asserting that communication often aims to share precise knowledge, which metaphorical statements cannot achieve as effectively as literal ones (
Allott and Textor 2022). They believe that, unlike relevance theory, which blurs the lines between literal and figurative meanings, a clear distinction is essential for clarity and effective knowledge transmission, particularly in fields like science and law. This latter idea will be further elaborated in the next section, where it will be discussed how the concepts of relation and distinction between literal and metaphorical language have a place in this article’s account of religious language.
I would like to shift our focus slightly to another important aspect that I believe is crucial for my arguments here: building upon these two theories, we may still be able to identify three distinct areas within language—strictly metaphorical, strictly literal, and those areas where the literal and metaphorical cannot be clearly distinguished—and this is possible when the context is well thought out. A criterion for identifying strictly literal statements is their absence of the processes of the concept narrowing–broadening (
Sperber and Wilson 2008, p. 91) or cross-domain metaphorical mapping (
Lakoff and Johnson 1999, pp. 12–13;
1980;
Kövecses and Benczes 2010;
Lakoff and Turner 1989). In areas of language beyond the strictly literal, providing a generalized categorical tool to distinguish between metaphors and literal content is challenging. However, it is also accepted that ordinary language is embodied, originating from perception and primary concrete concepts. While the origin of language is rooted in a literal foundation, later expansions push it far from that foundation toward more abstract and metaphorical forms.
11 Nevertheless, to classify certain usage as strictly metaphorical, careful consideration of the context in which the usage is employed is crucial.
This context sensitivity is particularly important in the explanation of religious language. It is reasonably argued that religious statements cannot be strictly literal due to the unique nature of the subject. However, this does not contradict the assertion that religious language can be literal in a broader sense. In this broader sense, the literality of religious language means that it is an example of conventional language.
12 However, we must be cognizant of the fact that conventional language literality
13 might not be fully understood if the strictly literal content is removed from it. On the other side, the irreducibility of metaphors is a subject of significant scholarly discussion, with many emphasizing its indispensable role in the discourse on the nature of religious language (
Davis 1989;
Soskice 1985,
2007;
Davidson 1978;
Searle 1969;
Stern 2000;
Alston 2019;
Camp 2006;
Cavell 2015;
Moran 1989). The argument posits that, given the divine’s transcendent nature, the language used to characterize the divine cannot be literal, as it does not directly refer to empirical concepts. Therefore, the context or the logic of God-talk permits the presence of both irreducible and regular metaphors in religious language.
If someone were to ask whether conventional language literality is the only form of literality applicable to religious language, I would respond, along with Allott and Textor, in the negative. However, the difference in my proposal lies in the fact that I give equal value to both metaphorical and literal elements. Their account, rather, gives more credit to literal language, as it is a form of language that can be understood independently; whereas, metaphorical language can only be comprehended in terms of the literal. They aver that if religious language is to convey information about the divine, it must incorporate a literal interpretation (
Allott and Textor 2022, p. 4). They support this claim by proposing that an “asymmetric dependency” between literal and metaphorical language serves as the criterion for distinguishing between them (
Allott and Textor 2022, p. 14).
Although the concept of asymmetric dependency seems to support the reducibility of metaphors, some recent scholars seek to use this distinctive feature to argue for the irreducibility of metaphors. Jacob Hesse, for instance, offers it as one of the criteria that can firmly establish the distinction. He states: “Whereas metaphorical interpretations necessarily presuppose the knowledge of the literal meanings of the expressions used metaphorically, literal interpretations do not presuppose the knowledge of possible metaphorical interpretations” (
Hesse 2023, p. 241). Another criterion, according to him, is the contextuality of metaphorical language, meaning that literal meanings can be acquired without reliance on specific linguistic and extralinguistic contexts; whereas, metaphorical interpretations are contingent upon these contexts. Hesse advances the argument by positing that at least a portion of religious language is inherently metaphorical. “Heuristic de re metaphors” are examples of irreducible metaphors that can be genuinely predicated of the divine. In his critique of Alston, he broadens the metaphorical realm by arguing for the plausibility of panmetaphoricism—the notion that all language about God is irreducibly metaphorical (
Hesse 2023, p. 240). Religious language, on this account, has a metaphorical core, where the divine is perceived through the lens of certain aspects that cannot be understood through solely literal expressions. This Wittgensteinian perspective highlights that the metaphorical nature of religious language renders it largely irreducible (
Van Herck 2013, p. 42).
In light of the discussion thus far, we have seen that the arguments concerning the nature of religious language, whether advanced by panmetaphoricists or by literalists (i.e., proponents of literal–metaphorical continuity), reflect certain convincing rational foundations. The pertinent question now is whether we have sufficient grounds to prioritize one perspective over the other. Should we abandon both views and lapse into quietism, or can we paradoxically assert that irreducible metaphors within religious language can coexist with literality? I propose a refined version of this perspective, which I term the dialectical interpretation of religious language. This approach aligns with an apophatic–cataphatic God-talk, while deliberately steering clear of quietism
5. Dialectical Dynamism and Concluding Remarks
To substantiate the proposal suggested at the close of the previous section, I would begin by highlighting the fact that the interpretation of religious language presents a particularly intricate challenge. Religious language, which seeks to characterize a transcendent entity cannot be paraphrased or reduced to strictly literal terms. This is because, while the divine is intimately connected to the world as its creator, it is not an object within the world but exists beyond it. The notion of irreducibly metaphorical religious language can create certain perplexities. While such a notion would highly respect divine transcendence, it encounters difficulties in explicating how language can articulate the relationship between the divine and the world, and how it is possible to know anything about God through inherently metaphorical language independently of a literal one.
Being cognizant of this problem, an intermediate position, as an alternative interpretation of via analogia, posits that analogical predication in theological language can adequately address both the divine relation and distinction. Nevertheless, this position has traditionally encountered criticism, such as the contention that the idea of analogy that is built on the concept of proportionality might imply that the meanings of terms predicated of God are used univocally, thereby rendering religious language, in the final analysis, literal. To mitigate this issue, some contemporary scholars (e.g., Burrell) propose that the concept of proportionality should be construed metaphorically, suggesting that all analogical language possesses a metaphorical core—an irreducible metaphoricity. The remaining challenge is to delineate how the metaphorical core is distinct from the peripheral layers of analogical language. Would this not constitute a variant of the so-called panmetaphoricism?
Stephen Mulhall is keenly aware of this challenge, acknowledging the tension between the cataphatic and apophatic aspects that contribute to the enigmatic nature of religious language. He would suggest that this mysterious nature does not entail that religious language is purely metaphorical; rather, it proposes that it incorporates both literal and metaphorical elements at once. By revisiting Wittgenstein’s philosophy and re-examining theological accounts of religious language, Mulhall sheds light on this mysterious aspect. He posits that religious language functions as a “riddle”, existing in a space between sense and nonsense (
Mulhall 2016, p. 3). Drawing from the distinction between promissory nonsense and hopeless nonsense, Mulhall explains that promissory nonsense, although not conventionally sensible, opens new avenues for understanding and reflection, especially in theological contexts. In contrast, hopeless nonsense fails to engage meaningfully with its subject and lacks the potential for deeper insight (
Mulhall 2016, p. 21).
Mulhall suggests that the “self-subverting” nature of religious language, as seen in promissory nonsense, transcends the typical meanings of words, engaging believers in an ongoing process of exploration and interpretation (
Mulhall 2016, p. 59). Religious language is a distinctive language game, where religious utterances, though seemingly nonsensical, hold profound significance within the life of faith. Mulhall’s account is particularly valuable, as it makes it possible for us to find confirmation of the mysterious yet genuine presence of a paradoxical nature, preserving both the literal and metaphorical dimensions of religious language—an idea further characterized in my approach outlined in what follows.
The dialectical position that I support would endeavor to capture the inherent paradoxicality at work in religious language. This entails accepting that the irreducible metaphorical and irreducible literal dimensions of religious language are not mutually exclusive. The dialectical approach distinguishes itself from previous proposals, including the notion that religious language is literal in one sense (conventional language literality, as introduced by cognitive theorists like Lakoff, Johnson, and Gibbs) and not literal in another sense (nonmetaphorical literality). Rather, it seeks to do justice to God-talk, where divine transcendence does not exclude divine immanence. While this position acknowledges that any language used in such a context cannot be nonproblematically applicable, it also recognizes the value of attempting to formulate new accounts of religious language rather than adopting a position of silence. Therefore, it suggests a departure from accounts that implicitly acknowledge, yet fail to explicitly pronounce, the paradoxical coexistence between literality and metaphoricity within the context of religious language. In contrast, religious language must utilize such enigmatic coexisting relationships to articulate the divine. On the one hand, it must employ strictly metaphorical language not only to negate the similarity between the divine and corporeal realms but also to highlight the limitations of our knowledge. On the other hand, it must also engage with a certain definition of literality in which the meanings of terms predicated of entities in both realms are highly connected, because the divine and the nondivine are intimately related. This form of paradoxicality is not illogical but rather coherent within the logical framework of God-talk,
14 and it is valuable in spiritual contexts. To illustrate the usefulness of engaging in such an intellectual practice, one might employ Wittgenstein’s ladder metaphor to suggest that religious language, by offering nonsensical and paradoxical horizons, challenges our language and understanding and enables them to ascend toward the realm of otherness. However, given the infinite nature of the entity to which religious language refers, the necessity for retaining the ladder persists, contrary to the notion of discarding it.
15Revisiting the interactionist notion of “tension” could help to more explicitly refer to the contrast between literal and metaphorical interpretations within the dialectical framework. The dialectical interpretation recognizes the merits of but advances beyond what Ricœur describes as “the logical structure of likeness”, characterized by the “tension between sameness and difference” (
Ricœur 1978, p. 148). This tension between opposing interpretations fosters the creativity and flexibility of language, enabling both language and thought to transform and approach the realm of the unthinkable and unknowable. Dialectical interpretation offers not only theoretical but also practical advantages. On the one hand, it functions as a form of contrastive analysis, employing analytical and grammatical tools to uncover the appropriate meanings of terms based on their use and context. On the other hand, it represents a continuously self-evolving practice, which undergoes transformation.
16 This account centers on a methodological strategy that I refer to as “analogical dialectics”, a concept that extends beyond the narrower scope of Tracy’s “negative dialectics”. In the context of negative dialectics, the dialectical enterprise engages in the so-called “hermeneutics of suspicion” to underscore divine transcendence and distinction, which are primarily metaphorical (see
Tracy 1983). Analogical dialectics, however, should be regarded as a deconstructive action within the constructive act of stretching toward the infinite. However, when the concept of negative dialectics is viewed through the lens of Hegel’s dynamic synthesis, one might be able to reach a constructive interpretation. Hegel’s concept of dynamic synthesis involves the resolution of contradictions through a process known as sublation (Aufhebung), wherein opposing elements are both negated and preserved, leading to a higher unity. This dialectical movement propels the development of concepts and reality, culminating in a more enriched understanding. Thus, each stage of synthesis incorporates and transcends previous contradictions, driving the progression toward absolute knowledge and the realization of the idea.
17 In this framework, a dialectical relationship between negative assertions would be understood as progressing toward affirmative content.
The manner in which Przywara engages with the concept of tension in the dialectical treatment of the notion of analogy is particularly inspiring. Przywara’s approach departs from Hegel’s dialectical synthesis.
18 While Hegel’s dynamic synthesis seeks to resolve contradictions through sublation, thereby integrating and transcending opposing elements into a higher unity, Przywara maintains the tension between opposites without resolving them. Another point of divergence is Przywara’s assertion that the analogy of being is a “suspended middle”, where the divine and the created order are held in a dynamic yet unresolved tension. This “unity-in-tension” in Przywara’s thought preserves the perpetual oscillation between divine transcendence and immanence, contrasting with Hegel’s aim of achieving absolute knowledge through the reconciliation of contradictions. Przywara plainly departs from Hegel by emphasizing the irreducible mystery of divine reality. What preserves the divine mystery and places it beyond complete human comprehension, as he asserts, is the radical otherness of the divine. Overall, while Hegel’s dialectic ultimately aims for a harmonious resolution, Przywara’s dialectical analogy seeks to preserve the “ever greater dissimilarity” within the way of analogy.
To sum up, contemporary scholarship supports the notion that language is, in principle, simultaneously literal and metaphorical, provided the context does not forbid it. It also supports the idea that, in the realm of religious language, this dual character possesses a tensive quality, wherein the metaphorical aspect shows a resisting force to literal interpretation, and vice versa. This study demonstrates that any coherent account of religious language must endorse a form of analogical dialectics in which negative analogy is integrated with positive dialectical analogy. This dialectical interaction between positive and negative poles suggests that God-talk, at its highest level, is a particular kind of no-talk, though at lower levels, it should aim for less-talk, so to speak. By embracing this approach, one can practically safeguard against the dangers of claiming to speak authoritatively for the divine, a tendency that has led to the historical and contemporary misuses of religious language. However, a comprehensive articulation of this idea extends beyond the scope of this article and represents a promising direction for future research. It is anticipated that future developments of this account of religious language will demonstrate its greater compatibility with the practice and discourse of religion in the public domain.