1To publish a book today that renews our approach to language, more than half a century after this question has been placed at the center of philosophical investigation, is no small achievement. Charles Taylor can be said to have accomplished this feat in The Language Animal. Published in 2016, this book completes a cycle that began with the 1978 article “Language and Human Nature,” and thus concludes nearly forty years of thinking on these issues. This book does not merely develop and illustrate the idea of an opposition between two paradigms for the understanding of language: the designative paradigm and the expressive paradigm respectively. It also takes a further step, which we will try to delineate. It endeavors to broaden our understanding of language by situating it in a certain continuity with other “expressive forms,” as Taylor calls them, that is to say, other ways by which human beings articulate and make manifest meanings: these range from expressive behaviors that are purely “enacted” by the body to the incarnations of meanings in arts that do not require verbal expression at all: dance, music, painting. Taylor thus proposes, in a gesture reminiscent of Cassirer’s, an understanding of verbal language that considers it in the context of human culture as a whole.

The two paradigms: The designative and the expressive

2To grasp the meaning of this broadening of our conception of language, we return to the same impulse as that of the 1978 article: the contrast between two approaches to language. The first type of conception, which Taylor associates with Hobbes, Locke, and Condillac, finds many echoes today in the post-Fregean philosophy of language. Here meaning is understood fundamentally as designation, with every word functioning on the model of a name. Each word is the sensible mark of an idea, and through the intermediary of this idea it relates to an extralinguistic reality. Locke writes that “the use then of words, is to be sensible marks of ideas; and the ideas they stand for, are their proper and immediate signification.” [1] Several fundamental consequences follow from this. First, a semantic atomism: the minimal unit of meaning is the word and language is basically an accumulation of words that could signify independently of each other. Second, a conception of thought as an essentially monological activity, of which language only ensures the communication to others. Nothing in principle prevents us from conceiving a language whose use would be essentially private. Shared language is only the fortuitous convergence of individual languages. As Locke points out, “when [he who speaks] represents to himself other men’s ideas, by some of his own, if he consent to give them the same names, that other men do, ‘tis still to his own ideas; to ideas that he has, and not to ideas that he has not.” [2] It follows that language has only an instrumental, not a constitutive role for thought itself. Language is endowed with a natural purpose, as Locke writes; it is an instrument for the communication of thoughts, and linguistic intelligence is only the highest form of instrumental ingenuity. Finally, owing to this essentially utilitarian function of verbal signs, everything that has to do with images and tropes in general can play only a peripheral role in language. The ideal of language is clarity, and metaphors should be excluded as much as possible, as Hobbes asserts. [3] Designative conceptions thus manifest a complete “anti-Cratylism” and are essentially “tropophobic.” [4]

3This designative paradigm received its canonical form in the theories of language of the classical and Enlightenment ages, during what Ian Hacking has called “the age of ideas,” [5] but it can also be found in more recent conceptions. The basic idea here is the same, that is, to explain linguistic meaning on the basis of the relation between elements of language (sentences) and an extralinguistic reality. Taylor qualifies these conceptions, in The Language Animal, as “enframing theories,” as “restrictive notions” of language. An illustration of this would be the Vienna Circle’s conception, according to which the meaning of a proposition is its method of verification, but this conception is also found more generally in theories that conceive meaning in terms of truth conditions. Here, language is conceived as a coding of thought allowing its communication, and meaning remains something relatively unmysterious, since it can be defined entirely in veridical terms. According to Davidson’s conception, I understand the meaning of a sentence when I am able to form a theory of my interlocutor’s language that allows me to specify the truth conditions of any sentence within that language, that is, to confront that language in its entirety with reality. According to such a conception, learning a language would be “learning to generate extensional truth conditions for its various depictive combinations” (LA, 262). The common feature of such conceptions, which are obviously much less naive than their classical counterparts, is that they continue to make the regimented language of logic—which makes it possible to adequately describe states of affairs and to determine in which cases the truth of a proposition is preserved within its inferences—a model for conceiving language in general. Of course, we are far from the naive atomism of meaning that presided over the designative theories of the “age of ideas.” Nevertheless, we find the idea that language is shared only in the second degree. By arguing that any situation of understanding a sentence of our language is in fact a situation of radical interpretation (like that which presides over the understanding of a totally unknown language), [6] Davidson questions the very idea of a shared language: each person is in a certain way the speaker of an idiolect within their own language. [7] According to such a conception, language remains endowed with a kind of natural purpose: that of communicating the truth about empirically testable states of affairs. As a result, all figures of speech and images have only a subordinate function in language: the meaning of a metaphorical expression is entirely determined by the combination of literal meanings it contains. [8] In Taylor’s view, it is this type of conception, which continues to take as its model the “purified” language of logic, that the second Wittgenstein first attacked.

4To escape from this dominant paradigm in post-Fregean thought and analytical philosophy in general, whose power Taylor understands well, we have to turn to a second family of theories, which Taylor refers to as “expressivist” or “constitutive” theories, and whose earliest manifestations go back to the Romantic period, in authors such as Herder, Hamann, and Humboldt. These theories stem from criticisms addressed to the designative paradigm. For Herder, for example, who proceeds with reference to Condillac, in the same way that Wittgenstein would later proceed with reference to what he calls “the Augustinian image of language,” the idea that the relation of the word to the thing, or of the word to the idea, provides the basis for the acquisition of language is an absurdity. In order to understand the designative relation, one must already understand what a name is, and thus master at least one language. Strictly speaking, only a second language can be (partly) learned by matching certain names to certain objects. What expressivist conceptions reject, therefore, is in the first place the alignment of words in general with names and the semantic atomism that follows from it. Any word has meaning only against the background of a language taken as a whole. This explains the organic metaphor that these authors use: in a language, no word has meaning independently of the others, and independently of the way in which sentences as a whole construct meaning. Of course, such a semantic holism does not imply that we cannot also assign meanings to words dissociated from their context of use, as a dictionary does; but it refuses to conceive the meaning of a sentence as a simple combination of the meaning of its elements. The use of a dictionary is always a secondary operation; it presupposes at least a partial mastery of the language (or at the very least of another language). As a result, language is not the secondary vehicle of thought, a simple instrument at its service, but the living embodiment of it, and it therefore fulfills a constitutive role with respect to thought. There is no thought, in the human sense, without words. Language is Bildung, as Humboldt says, it is that in which humanity and thought itself take shape. Since any precedence of thought (or “ideas”) over language must be rejected, it follows that the idea of an essentially private language, of a language constructed by the individual as a simple encoding of their thought, is an absurdity: “bound and dependent on the nations to which they belong,” Humboldt asserts, [9] it is always the language of a given community. Finally, language does not proceed from an intellectual operation to share and communicate thoughts that would already be formed independently of it; it informs our feelings and our entire sensibility. Expressivist theories of language do not only contest the idea that we can conceive of linguistic meaning on the model of the relation between words and an extralinguistic reality; they go so far as to question the idea of an extralinguistic reality that would be what it is for beings of language. It confers on language a constitutive dimension for the experience of the world itself. Language, as Humboldt states, is a Weltanschauung. Indeed, according to these conceptions, the characteristic of a verbal expression is not only that it manifests outwardly a thought to make it accessible to others. The expression does not only communicate, it “constitutes what it expresses.” [10]

5In this respect, it is appropriate to distinguish two concepts of expression. According to the first, expression is only contingent with respect to the expressed content; according to the second, it is necessary for its very existence. It is this second concept of expression that serves as a guideline for expressivist theories. The result is that “what expression manifests can only be manifested in expression.” [11] This is the case not only for a poem, but also for other non-verbal expressions—musical, pictorial, or others. Such a conception places the literary image or the rhetorical trope at the center of our understanding of language, since these incarnations of meaning cannot be understood independently of the figure in which they take shape, as mere combinations of literal meanings. The expression is constitutive of what it expresses, and which cannot be understood independently of it, in accordance with a conception analogous to that which Merleau-Ponty had already formulated, and to which Taylor is here rather close: as Merleau-Ponty asserts in the Phenomenology of Perception, “the expressive operation actualizes or accomplishes the signification and is not merely a matter of translating it. But despite appearances, the same is true for the expression of thoughts by speech. Thought is not “inner,” nor does it exist outside the world and outside of words.” [12]

6We can thus see what separates the two approaches to language. For the constitutive theories, language does not limit itself to taking place within a reality and a set of human practices that would precede their expression and would be what they are independently of the latter. Language creates a new dimension in the world that is only made possible by language. It allows a new type of consciousness of the world and of ourselves. “If language serves to express/realize a new kind of awareness,” Taylor writes, “then it may not only make possible a new awareness of things, an ability to describe them; but also new ways of feeling, of responding to things.” [13] In other words, “we might say: language transforms our world, using this last word in a clearly Heidegger-derived sense. That is, we are talking not of the cosmos out there, [...] but of the world of our involvements, including all the things they incorporate in their meaning for us” (LA, 37). Thus, we can say that language contributes to shaping the essentially human meanings that populate our world: we can only experience certain feelings, such as indignation, because we possess a rich moral vocabulary, bringing into play fault and innocence, the just and the unjust, the noble and the ignoble, etc. Contrary to the enframing theories that claim to explain linguistic meaning on the basis of the relation of signs to an extralinguistic reality, a constitutive theory “gives us a picture of language as making possible new purposes, new levels of behavior, new meanings, and hence is not explicable within a framework of human life conceived without language.” [14] These theories are not only found in pre-Romantic and Romantic thought, but their influence can be traced back to conceptions like those of Heidegger and Gadamer. It is in this “constitutive” perspective that we can understand, says Taylor, the Heideggerian image of language as a “house of being,” or the Gadamerian affirmation of an essential Sprachlichkeit [language dimension] of our experience of the world. [15]

7Here, however, a difficulty arises. If language possesses a constitutive dimension in the strong sense that Taylor intends to give to this notion, if it entirely informs our very experience of the world, how can we reconcile this assertion with the realism that Taylor also defends—with his moral realism, of course, but also his epistemological realism, as he expressed it in particular in a book co-written with Hubert Dreyfus, Retrieving Realism? In fact, from the historical point of view the most resolute formulations of the expressivist paradigm take place in the wake of Kantian transcendental idealism, as is the case, for example, for Humboldt’s theory. The inner form [innere Sprachform] of language is what, in this conception, plays the role that Kant assigned to the categories of understanding for shaping our sensible receptivity through the spontaneity of the mind. Certainly, Humboldt inflects Kantian transcendental idealism in the direction of a linguistic idealism, since it is no longer individual understanding, in its union with sensibility, that allows the constitution of objectivity in general, but rather a particular language as the intellectual production of a people. Nevertheless, the ideality of language is at the basis of a shaping of our experience, concealing things as they exist independently of the mind and of ourselves. Language thus becomes a filter or a medium of which the mind remains captive. Should we not note the same fundamentally anti-realist tendency in the Heideggerian image that, as Taylor writes, consists in characterizing “the clearing [of being] as the space opened by expression”? [16] Is it not then language that makes possible our very access to things and is placed in a quasi-transcendental position? Is this not what emerges from passages such as this one:

8

Because language [Sprache] is the house of being, we access being only by constantly passing through this dwelling house. When we go to the fountain, when we cross the forest, we are always already crossing the word “fountain,” the word “forest,” even if we do not utter these words, even if we do not think of language. [17]

9Should we not conclude, in the same way, that there is an affinity between the idealist tradition and Gadamer’s affirmation of the essentially linguistic character of our experience of the world? For, as Gadamer specifies, “the idealization of language [. . . ] is already present in any acquisition of experience.” [18] And, consequently, “the world [is world] only insofar as it comes into language.” [19] Gadamer moves, at least in such passages as these, beyond the widely shared idea that the possession of language contributes to orienting our apprehension of the world and to modifying the experience we have of it; he seems rather to maintain that our experience of things is determined by language throughout. But then how could our particular language be anything other than a new interface between the world and ourselves? In fact, the whole point of Taylor and Dreyfus’s work on rethinking realism is precisely to free us from what they call the “mediational epistemology” inherited from modern philosophy. It is therefore to this critique that we should turn.

The rejection of mediational epistemology

10In order to address this question of the compatibility of the constitutive conception of language with the direct realism defended in Retrieving Realism—a question also raised by Christoph Demmerling [20]—we will provide a quick overview of the latter work. The argumentative strategy of Taylor and Dreyfus consists in showing that it is not enough to reject the classical theory of ideas in order to escape the mediational image of knowledge, and that when authors such as Rorty, Davidson, or McDowell replace ideas by beliefs by arguing that “nothing can count as a reason for holding a belief except another belief,” [21] and that, consequently, “there is no way to get outside our beliefs and our language so as to find some test other than coherence,” [22] they only perpetuate the same image of confinement within the space of thought that makes beliefs the only “inhabitants” of the space of reasons. But this apparent truism of confinement in the circle of our beliefs and of the impossibility of confronting these beliefs with reality must be purely and simply rejected. In reality, we are perfectly capable of making sense of the idea of a direct grasp on reality by means of our perception, if we understand this perception as a bodily capacity that we exercise with a certain competence, answering to certain criteria. Let us suppose that I ask Johnny to go and check if a picture hanging on the wall is crooked. Provided that Johnny competently exercises his visual abilities under optimal conditions, that he places himself in the right place, looks in the right direction, focuses on the object that requires his attention, waits until the lighting is sufficient, etc., the knowledge that he draws from this is not that the picture seems to be hanging crooked, such that it would still be necessary to check that this is in fact the case; the knowledge he draws from it is that the picture is crooked. For there is, in truth, no other way for him to verify this point than to exercise his perceptive abilities under optimal conditions. If this does not persuade him, if these criteria are not sufficient, then nothing in reality can ever persuade him. It is therefore quite possible to make sense of the idea that we measure our beliefs directly against reality by using our skills as perceiving subjects—without having to enter into absurd skeptical scenarios according to which we would have to “get out of our skin,” as Rorty says, in order to confront our beliefs with the facts. The conclusion that follows from this is that the apparent “Davidson-Rorty truism is false,” [23] and that what ensures us a direct “grasp” on the world, to use Merleau-Ponty’s vocabulary, are our coping practices, our practices of bodily commerce with the world. And these practices allow us to relate to the world not through any mental interfaces (explicit representations or beliefs) but by relying on an implicit background knowledge that comes to us from being situated bodily in this world. We have to be immersed in the world through our body in order to be able to represent (including by means of beliefs) this or that aspect of it. It is because we are connected to the world by an “umbilical cord,” as Merleau-Ponty writes, the one that guarantees our incarnated perception, that the very idea of a perception that would not open onto any world is a nonsense. As Merleau-Ponty writes in Phenomenology of Perception: “we must not wonder if we truly perceive a world; rather, we must say: the world is what we perceive.” [24] It is also, obviously, this presence in the world unmediated by representations that presides over the Heideggerian notion of being-in-the-world.

11Our bodily engagement with the world thus ensures our immemorial contact with it without intermediaries of any kind, and is at the basis of the direct realism that Dreyfus and Taylor assert. It is not by chance that such a realism must then make room for the recognition of a prelinguistic dimension of our experience. This is what is specified in chapter 3 of Retrieving Realism, entitled “Contact Theory: The Place of the Preconceptual.” Most of the capacities by which we cope with the world are capacities that do not involve conceptually articulated thought and judgment at any point. These preconceptual coping skills place us in a relationship with the world that is prior to any shaping of it by language. Taylor and Dreyfus illustrate this point with the following example:

12

Take the prelinguistic. Of course, much of our human world is already linguistically articulated. We move among chairs and tables, outdoors, into cars, towards buildings and laboratories. But it is never completely so, in the sense that the range of articulacy can be extended. As a boy I go with my dog every day to explore the woods yonder. We have to cross a stream, and we both hop across on some conveniently placed rocks. I don’t have a word for these; I don’t even feel the need for one. We both just hop across on our way to the inviting, mysterious woods. Then my older cousin comes to visit, sees the woods and the stream, wants to cross, and asks me if there are any “stepping stones.” Because “stone” and “step” are already in my vocabulary, I get at once his meaning [. . .]. These hitherto mute facilitators have entered the linguistic dimension for me thanks to my cousin. [25]

13Direct realism requires a recognition of the prelinguistic dimension of at least part of the meanings that populate our primordial experience. Should it not then impose a limit on the idea of language as a “clearing of being,” or on the constitutive powers of language for the shaping of our being-in-the-world itself?

The radicalization of Taylor’s theory of language

14Rather than trying to answer this question directly, and with a view to doing so, we would first like to consider the way in which The Language Animal radicalizes the intuitions already present in “Language and Human Nature” and the texts of the same period. We propose to identify at least three points on which Taylor refines and deepens his own positions.

15(1) First of all, Taylor radicalizes Frege’s contextual principle beyond even the extensions that have been proposed so far. For Frege, as we know, “we must never try to define the meaning of a word in isolation, but only as it is used in its context,” [26] and this context is that of the proposition. Wittgenstein further extended this semantic holism by questioning the primacy of the assertion and by making room for the different activities that we can accomplish thanks to language. Words have meaning only within language games, and the latter are possible only within forms of life that are not exclusively verbal in nature. This is what makes Wittgenstein say that if a lion could speak, we would not understand it. It is then all human practices and ways of relating to the world that contribute to the meaning of what is said. Taylor agrees with this view: “the basic thesis of this book is that language can only be understood if we understand its constitutive role in human life” (LA, 261). But he goes in a sense further than Wittgenstein by determining further the nature of this background that is necessary to the understanding of verbal language and by including in it all the other symbolic or expressive forms that maintain an obvious affinity with verbal forms, while also being distinguishable from them: gesture, facial expression, but also dance, music, mime, pictorial forms. It is necessary to go beyond a simple pragmatics, by asserting that language serves more purposes than just doing, it is one of the ways in which we can articulate meaning on a par with other non-verbal forms of articulation. It is then possible to reformulate Wittgenstein’s expression, which is pivotal to a semantic holism: language must be placed “in a broader context, that of the range of symbolic forms which run from pure assertoric to pure disclosive.” [27] Thus, even Brandom’s pragmatics, which also seeks to broaden the contextual principle beyond isolated language games, by taking assertion and inference as its guiding thread, continues to provide an excessively impoverished view of language, one that is too exclusively centered around the exchange of information and justification, whereas narration, myth, fable, and the poem constitute equally paradigmatic and primordial manifestations of what speaking means for a human being. The pragmatic extension of the contextual principle is still insufficient: “the factual-practical can’t be self-sufficient. Our ability to operate with this family of language games depends on our operating in the whole range of symbolic forms.” [28]

16(2) The extension of the contextual principle to the totality of the modes of articulation of meaning logically leads to Taylor’s second decisive innovation: an extension of the very concept of “language.” What we say is never everything. The gestures with which we accompany speech, and even the gestures that are themselves speech, are part of this articulation of meaning that goes beyond the forms of verbal language and without which language itself would be incapable of meaning. As Taylor explains, “language as speech can only exist in symbiosis with various forms of embodied action—gesture, enactment—as well as other symbolic forms, music, dance, poetry, and other modes of artistic expression. The ‘country’ of language goes way beyond the ‘province’ of information-encoding, important as this is” (LA, 99). To understand this extension that it is useful to attribute to “language” in general so as to be able to delimit the province of verbal language, Taylor claims that the common feature of all “language” is its power of articulation and illumination—observing, in passing, the Indo-European etymology of a whole family of terms referring to speech, the radical phan- (*bha-, *bhô-) which means “to illuminate, to shine,” but also “to explain, to speak,” and which is found in ancient Greek in phasis, declaration, appearance, or in apophanein, to declare, but also to make appear. [29] Human language should not be understood as being mainly composed of assertions, nor even of speech acts; it fulfills a function that it shares with the work of art: it is “disclosive,” it can “make something manifest by articulating it.” [30] What is true of verbal language is also true of other modalities of language, such as music. Chopin’s Fantaisie-Impromptu articulates meanings in this sense, it possesses a “depictive” dimension, even if the meanings it articulates—an indefinable sadness or nostalgia—are impossible to translate entirely into words and possess, in this sense, something that would be “incommunicable” in other ways (LA, ch. 6). A “ladder of articulacy” is thus set up that goes from the simply “enacted” meaning—here, Taylor takes the example of the way of comporting oneself, of expressing oneself, of walking with a swagger or of dressing like a biker, which we can try to capture by means of the word “machismo”—to disclosing in the form of assertion or description, and finally, to disclosing in the form of a symbol, which is characteristic in particular of works of art. To perceive the continuity between these different ways of articulating a content by figuring it allows us to put the figurative dimension of language back at the heart of our understanding of it, which was also in many ways Gadamer’s project. [31]

17(3) Finally—and this is perhaps the most remarkable of these three refinements—Taylor’s conception, by resituating language in the narrow sense in the context of human language in its most extended sense, also aims to restore to the expressive body and gesture their true role in the intelligibility of language (in the narrow sense). While one could detect in the first two refinements an inheritance from hermeneutic thought (Heidegger, Gadamer, but also Ricœur), this last feature of Taylor’s theory of language is rather indebted to Merleau-Ponty. Taylor deliberately cites Merleau-Ponty’s texts in which the latter establishes a continuity between language and expressive gesture (LA, 29-30), and affirms that “speech is a genuine gesture” or that “the linguistic gesture, like all others, sketches out its own sense.” [32] It is not so much a question of widening our conception of language as of being attentive, in speech, to all this preverbal, gestural, and perceptive infrastructure which is like its nourishing soil and without which it could not blossom. A gesture, in fact, unites motricity with meaning. It is not a simple movement in space, but a movement which supports its own meaning, which secretes a sense from which it cannot be separated: a gesture of anger does not make us think of anger, it is anger itself. In the perception of this gesture as anger, form and meaning are one. It is because a meaning emerges amid the intricacies of sensibility, on the level of perception itself, that this meaning can be taken up, articulated, and reshaped by language. This third characteristic makes it possible to remove semantics itself from an “intellectualist” approach, as Merleau-Ponty would say, or from what Charles Taylor refers to as “the old disengaged epistemology” (LA, 31), which thinks of the speaker as an acosmic subject manipulating signs that are transparent to him in order to encode simple mental representations. This disintellectualization of meaning has no real equivalent in the hermeneutic paradigm of Heidegger and Gadamer; for what is lacking in their approaches to language is precisely this attention to the continuity of language with gesture and perception. As Taylor points out:

18

the linguistic capacity is essentially more than an intellectual one; it is embodied: in enacted meanings, in artistic portrayals, in metaphors which draw on embodied experience, and also in the iconic gestural portrayal which accompanies everyday speech, not to mention the ubiquity of “body language”—tone of voice, emphasis, expressive gestures, stances of intimacy, of aloofness—which surround ordinary discourse. (LA, 333)

Beyond the linguistic turn?

19Surely it is precisely this third characteristic of the refinement and deepening carried out by The Language Animal in the formulation of the expressivist theory that makes it possible to find a solution to the problem we have raised, or at least to clarify the difficulty differently? The tension between the rejection of the mediational epistemology, based on the idea of a direct bodily “grasp” on the world, and the constitutive theory of language which makes language into a condition of possibility for the very appearance of things, and thus a new interface between the world and us, tends to vanish as soon as we understand that language extends and metamorphoses at the same time a meaning that is indissociably enacted and perceived, a meaning that emerges at the level of our sensibility in relation to the world—the level of what Merleau-Ponty, after Husserl, called “silent meanings.” Because language is embodied, the linguistic semantics rests on a “natural” semantics without which language could mean nothing at all, and this natural semantics takes place at the level of our primordial bodily interaction with the world, unmediated by representations. It is because language is in continuity with other already signifying modalities of our relation to the world, at the very level of our embodied sensibility, that this language, although it certainly modifies our experience of things and our sensibility, does not constitute a mental prison at all. The emotional, aesthetic, and cultural meanings that are only made possible by speech, and in relation to which speech plays a constitutive role, would not be possible, in their turn, without our direct bodily access to a reality that is independent both of ourselves and of language—an access without which language itself would remain unintelligible. This is what emerges, in particular, from the whole discussion of the Sapir-Whorf hypothesis on the role of language for the very perception of colors, developed in chapter 9 of The Language Animal. Sapir and Whorf are representatives of a neo-Humboldtian semantics that holds that the language we speak shapes our experience throughand through, and they thus arrive at the assertion that “the worlds in which different societies live are distinct worlds, not merely the same world with different labels attached.” [33] Taylor refuses such a position, by stressing that although our vocabulary of colors, for example, can make us more attentive to certain chromatic nuances than to others, for example to "Sienna ochre," if we have this vocabulary, it does not determine our perception of the colors, which remains governed by the same fundamental invariants as for other cultures. Taylor’s insistence on the importance of Tomasello’s “joint attention,” i.e. of a pre-language basis for language acquisition (LA, 55-58), or on the need for perceptual “templates” (LA, ch. 5)—we learn the names of animals by starting with “tiger” and “lion,” not with “feline”—tends in the same direction. Our understanding of language itself is not purely linguistic. But then, we must abandon at least part of the constitutive view as it was previously conceived, that is, the linguistic idealism that underlies many of its formulations. We must assert that the partially constitutive role of language for our very experience of things does not exclude the assertion in principle that our embodied perception provides us with an access to things that is not always already fully made possible by language—nor mediated by anything else. Language, like all symbolism, belongs to the order of representation, but our primordial access to the world is not of a representational order.

20It is thus in his Merleau-Pontian phenomenology of the expressive gesture that Taylor finds the means to escape the specter of linguistic idealism that impedes many constitutive theories, including, no doubt, those of the second Heidegger and of Gadamer. It is by allowing a place for the non-verbal within language itself, continuously underpinning and accompanying language, that he is able to escape the illusion of language as a prison or Weltanschauung. The question then arises as to whether his contribution to these questions should be situated in the line of the linguistic turn—to which one can still in some way link Heidegger and Gadamer—or whether it instead moves beyond this linguistic turn. Taylor himself seems to agree with the former solution. He affirms that it is “absolutely right to consider [his] work as part of the (many-stranded) ‘linguistic turn’ of the second half of the twentieth century, which englobes analytic philosophy, as well as ‘post-structuralism,’ and the Heideggerian-Gadamerian ways of proceeding.” [34] But is this really true? The linguistic turn, at least in its analytic versions, was mainly based on the idea that language was something relatively clear, because it was accessible to analysis (in different possible senses), contrary to other domains, for example that of experience, and that it was possible to solve (or even dissolve) traditional philosophical problems by transposing them to this terrain and reformulating them in “linguistic” terms. But not only does Taylor insist on the obscurity of language, even on its “mystery,” [35] but he assigns the origin of this mystery to the anchoring of language in our bodily motricity and our sensible experience of the world. We might conclude that one of the major merits of Taylor’s great book on language is precisely, by resituating our understanding of verbal language in the context of other expressive forms, but even more by reinscribing verbal language in its preverbal infrastructure, to deepen the linguistic turn up to the point where it must metamorphose into an experiential turn; that is to say, to put language, as an expressive form, back into our being-in-the-human-world in its entirety.