In his book The Language Animal Charles Taylor presents a thoughtful criticism of the view of language that is widespread within the mainstream of analytic philosophy. He refers, among others, to Johann Gottfried Herder and Wilhelm von Humboldt, to phenomenology and hermeneutics, in order to connect philosophy of language and philosophical anthropology. This is an important undertaking because Taylor tries to correct many of the prejudices that guide conventional philosophy of language primarily in the English-speaking world. Speaking a language means participating in a form of life. In order to understand a language, one must get the lives of its speakers into view, not only their linguistic lives but also their whole way of dealing with things and the world. This is, I think, what was meant by Wittgenstein when he wrote: “If a lion could talk, we could not understand him.” Footnote 1
First, I would like to stress my great agreement with many of the claims defended by Taylor in The Language Animal and I would like to express my general endorsement of the whole approach presented there. The meaning of linguistic utterances does not exist independently of the social and interpersonal context in which they are used. The representational picture of language that is defended by many post-Fregean philosophers of language and mind is deeply misleading. Their view of language is shaped by a series of requirements concerning logical and scientific usages of linguistic signs. They cannot, therefore, come to see the strong relation between understanding a language and understanding the human form of life. This is the picture presented by Taylor, following in the footsteps of Herder and Humboldt. I presented a similar view following not Herder but Heidegger and Hans-Georg Gadamer in my book Sinn, Bedeutung, Verstehen. When I read The Language Animal I remembered that at the time I was writing my book I was also deeply influenced by some of Taylor’s earlier works. Footnote 2 I say this in order to make clear from the beginning that my comments will not be very critical. Rather, I would like to inquire into the details of some of his points. I will first talk about the constitutive view of language (I) and then take the concept of a concept into consideration (II). A third remark concerns the relation between language and feelings or emotions (III). While all these points may be linked to one another, I will treat them each separately. Each part of the paper identifies some open questions in Taylor’s proposal.
I. Language
My first point concerns the constitutive view of language. Language creates, figures, and shapes human experience or what we call ‘world’ in the broadest possible sense of the term (including the whole realm of non-sensual entities like numbers, norms, capacities of the mind, etc.). Taylor’s book is full of convincing examples that serve to drive this point home. My question is: How can the constitutive view be reconciled with realism? In the last book Taylor wrote prior to The Language Animal (co-authored with Hubert Dreyfus), readers are confronted with a criticism of the “mediational view.” Footnote 3 Taylor and Dreyfus criticize the idea that human beings experience their world only through something else: through internal representations or ideas (Descartes), through language (Frege), through surface irritations (Quine), or through brain states (contemporary materialism). One aim of this work was to overcome the mediational view and to argue for the claim that there is a direct relation between mind, body, and the world. Many philosophers within the tradition of phenomenology, such as Heidegger and Merleau-Ponty, tried to establish the view that human beings are directly connected to the world. Footnote 4 But the constitutive view of language, which Taylor endorses in his writings on the philosophy of language, is an exemplary form of the mediational view. To put it bluntly, either language shapes our way of experiencing the world, and then seems to mediate the relation between man and world, or human beings are directly related to the world. In the former case, it is a little obscure exactly what ‘constitutive view of language’ means. So, my question is: How would Taylor describe the relation between the constitutive view and the mediational view? In other words: Isn’t there a conceptual tension between the Herder-line and the Heidegger-Merleau-Ponty-line in the new book? How is it possible to reconcile the constitutive view with direct realism?
I think the tension becomes particularly obvious in one of the last chapters of The Language Animal, where Taylor tries to weaken and to defend the Sapir-Whorf-Hypothesis at the same time. Footnote 5 Linguistic relativism is clearly committed to the mediational view. On the other hand, Taylor argues against the idea of a fundamental incommensurability between different languages that is often associated with the Sapir-Whorf hypothesis. He refutes the idea of incommensurability by claiming that differences in lexicon (i.e., colour vocabularies) and grammar only “require that we pay attention to different things,” but this does not mean living in different worlds. Footnote 6 This is a very weak reading of the Sapir-Whorf hypothesis. Of course, as Taylor also tells us, there are other levels of language. Language is not only used to describe our natural or artificial surroundings and to make their features salient. It is also used to say something about the world in general and to present a view of the whole. At this level, one can find greater differences than at the level of a description of our surroundings. Concepts of space and time are used, and words are needed to refer to abstract or non-sensual things. Taylor admits that there could be great dissimilarities between the cosmic views entailed in different languages and he defends a stronger version of the Sapir-Whorf hypothesis now. He is even willing to accept the talk of incommensurability in these cases: “So while the response to the different coding strategies discussed above might just be that the differences are minor and undramatic, our reply to the deep and baffling differences between English and Hopi conceptions of time could just be: ‘so what?’” Footnote 7 But what does this “so what” mean? Are differences between our statements about the world just not as important as we usually think? Or is it rather the case that incommensurability doesn’t matter, because one can always learn to see something new in the world? Perhaps Taylor’s relaxed attitude towards the issue is justified. But if language creates more or less great differences between worldviews, then language mediates our experience, at least on the relevant levels, through and through. In sum, I would like to point out that two questions are open: How is it possible to reconcile the criticism of the mediational view with the defence of the constitutive view? And is there a sense in which the constitutive view can be classified as realistic?
II. Concepts
I would like to begin with one of Taylor’s favourite examples of embodied meaning in his book (that is meaning without words). Taylor describes the case of a leather-jacketed motorcyclist, whose behaviour and body language says something. Footnote 8 It reflects a way of dealing with things and the world. One can use the linguistic designation or the concept ‘macho’ in order to refer to his outlook and his behaviour. Does the motorcyclist have the concept ‘macho’ even if he has never heard or used the word? If this is the case, it may be possible to possess a concept without possessing a word. Is it also possible for an animal to possess a concept? Has an animal using a stick as a tool the concept of a tool? Of course, this case is different from the case of the motorcyclist but in both cases concepts are involved without using or even having words. Perhaps the discussion of and confrontation with such cases could be a good starting point to investigate the relations between concepts and language.
I didn’t detect much talk about concepts in The Language Animal and I wonder if, according to Taylor, the mastering of a language is a necessary condition for having concepts. I was directed to this question because of the distinction Taylor makes between different levels or kinds of meaning: biological meaning and metabiological or human meaning. Footnote 9 But another distinction is also in play there, namely that between the non-linguistic meaning of things or matters of facts, on the one hand, and linguistic meaning, the meaning of signs, on the other hand. For the sake of clarity, let us call the first kind of meaning ‘significance’ while for the second kind we can simply use terms like (linguistic) ‘meaning’ or ‘sense.’ For living beings everything in the world counts, everything could be of importance in the sense that it could be attractive or repulsive. Nearly everything in the world could be a matter of concern. Things in the world and matters of fact are not simply there; they have significance for living beings. A bone is significant for a dog. The dog takes it as a toy. The bone offers an ‘affordance’ to play with it—to use a term coined by James J. Gibson. Footnote 10 A cat is significant for a mouse. The cat is taken or perceived as a predator. It is an affordance to flee (I know it sounds strange to indicate a threat as an ‘affordance’ but I think we can take the term in a general sense indicating a challenge to act in a certain way). For us, human beings chairs, bikes, and books are affordances. We take them as opportunities to sit down, to ride, or to read.
If reacting to an affordance means taking a bit of reality as something, then is a concept involved? Pace Dreyfus, perhaps also contrary to Taylor, but in accordance with Heidegger, I would like to say that concepts are involved in the described cases. The application of a concept is an ability to take something as something. In § 32 of Being and Time, Heidegger writes: “The ‘as’ constitutes the structure of the explicitness of what is understood; it constitutes the interpretation. The circumspect, interpretive association with what is at hand in the surrounding world which ‘sees’ this as a table, a door, a car, a bridge does not necessarily already have to analyze what is circumspectly interpreted in a particular statement. Any simple and prepredicative seeing of what is at hand is in itself already understanding and interpretative.” Footnote 11 Heidegger indicates that when we deal with things in the world the structure of ‘taking something as something’ is always in play. That means that things and states of affairs are experienced always as classified in different respects. Does this classification require language? Does it involve statements or judgements? Does it involve concepts? Heidegger draws our attention to the fact that taking or perceiving something as something is a kind of interpretation, which leads to understanding. One question concerns the meaning of the term ‘pre-predicative’ in the quoted passage. Does the term refer to something beyond a linguistic statement but within the realm of concepts? Or does it refer to something pre-conceptual? To understand something pre-predicative means to have a concept without necessarily having a language. I take Heidegger’s distinction between a hermeneutical ‘as’ (understanding without a statement) and an apophantical ‘as’ (understanding by a statement and expressed by a statement) as a clue and a possibility for introducing non-linguistic concepts. Of course, this issue is controversial. Footnote 12 Be that as it may.
Certainly, it doesn’t suffice to discriminate things. One must be able to identify and re-identify them, to correct identification errors, etc. Concept use is related to correctness conditions. But this doesn’t necessarily mean that a language is involved. If someone uses a stick as a tool, he or it must have a practical concept of this tool. So, we must distinguish practical concepts from linguistic concepts and then investigate the conditions for concept-use in a detailed manner. Furthermore, we must study the intimate connections between the non-linguistic and the linguistic applications of concepts. Of course, one also must take into consideration the similarities and the differences between linguistic concept users and non-linguistic concept users. My claim is that linguistic meaning grows out of non-linguistic significance. High-level and linguistically mediated processes of understanding, which are meaningful, emerge from low-level and non-linguistic, bodily-sensual mediated interactions with the environment, which constitute significance. Meaning and significance must be understood as mutually constitutive phenomena, whose interrelation can be shown and made explicit on the basis of an analysis of the preconditions of linguistic understanding. The relevance of concepts for linguistic and non-linguistic understanding is especially important in this regard. As constituent parts of thoughts, concepts are considered to be essential for both kinds of understanding. The notion of a practical concept can be used to conceive of instrumental behaviour and perceptual experience as non-linguistic attachments to the world, which are nevertheless ways of understanding.
Taylor pursues a similar idea by arguing for the claim that “knowing our way around is a kind of understanding.” Footnote 13 Living beings, not only humans, have a sense of their surroundings, of things and objects that can be used in a certain way. In the case of human beings, this sense underpins the grip of linguistic meanings. Taylor talks about a “penumbra of meanings which surrounds our grasp of familiar things” Footnote 14 in order to frame this point. Unfortunately, Taylor doesn’t talk about concepts and I wonder why not. It remains an open question in his book as to how concepts are related to language. Is the fundamental sense of meanings, which is shared by humans and higher animals, conceptual or is it pre-conceptual?
III. Emotions
In Taylor’s book, there are many passages dealing with feelings and emotions. Of course, it won’t be possible to discuss all aspects of the comprehensive image presented. If I am not mistaken, Taylor considers various questions but it is not clear exactly how they are interrelated. First, there is the idea that meaning is felt. Footnote 15 Then there is a second idea, that feelings and emotions have meaning or—as Taylor puts it—they are “as charged with meaning.” Footnote 16 And finally the question of “how language can help constitute our emotions” is discussed. Footnote 17 How are these three points interrelated? Or is there no relation? Even if there is a relation, I would like to consider two of the points separately. Let me raise two quick questions:
-
(1) What does it mean that meaning is felt? Emotions are felt and they have a meaning as well. The meaning of emotions can often be described with the help of a that-clause. Shame, for instance, as an emotion indicates that someone has violated norms that are important to them. In this sense, shame has a meaning. It is about something and concerns the violation of acknowledged norms. Shame is felt. But the meaning of shame isn’t felt. The meaning is understood. Perhaps Taylor’s aim was to defend the idea that the meaning of verbal and non-verbal expressions can be felt. But this idea isn’t clear to me either. In this case too, I would say: meanings are understood. Certainly, processes of understanding can be more or less explicit; they can be intuitive. But a less explicit, intuitive understanding isn’t identical with a feeling.
-
(2) Language can constitute, clarify, and enhance feelings. However, in the easiest case, language is used simply to report feelings or emotions. Sometimes the searching for and finding of words is a way of becoming aware of, or making explicit, a lived experience. Watching the lifestyle of the rich and the beautiful, I feel a diffuse kind of anger. I start to think about it and I am convinced to feel indignation in the sense of a moral emotion. In the course of thinking about the case again and again something can change. Suddenly I have to admit that I am simply consumed by envy. This is an example of the clarification of an emotion. But it is not the case that language creates the emotion as a whole. There is some kind of emotion before my words do their work. The view that language, and expression more generally, is constitutive of one’s emotion is not incompatible with the view that emotions can be individuated before their expression and that linguistic acts merely report or describe them. The question is: At what point exactly do linguistic affairs come into play?
I propose that humans are creatures who live their lives entangled in a web of language. The thesis that I endorse emphatically is: only through linguistic articulation do states such as thoughts or emotions take shape. This does not mean that there would be no reactions to occurrences in the environment or felt experience without linguistic abilities. It simply means that emotions only take shape through linguistic articulation. Only then can they be placed in a space of understanding, in the context of which they can then be individuated by being embedded in inferential relations characteristic of thoughts and actions.
Let me illustrate this by way of an example. Aggressive affects can be found both in creatures capable and incapable of language and it is beyond doubt that not only adult humans but also animals are able to behave aggressively. Complex distinctions of aggression into rage, irritation, anger, envy, and indignation are, however, only opened up by linguistic means. The kind of aggression that sets in if one is, for instance, prevented from following one’s own plans (anger) is different from the aggression that sets in if one is of the opinion that someone, without deserving it, has something that one has not (envy). A distinction such as that between anger and envy, or any other aggressive affect, presupposes differentiated attitudes to the world, to the properties and actions of other human beings and these attitudes cannot, in all likelihood, be had without a language by means of which propositionally structured thought can be formulated.
Through the ability to articulate something in an explicit manner, multifaceted dimensions of the significance of situations can be distinguished, as the example of aggression shows. The anger at being hindered from following one’s own plans (for instance, due to a delayed train that was supposed to take one to an important appointment) is a different kind of aggression than the indignation at someone else’s disregard for a norm (for instance, lack of loyalty in a co-worker). Such differentiations only become possible through the fact that the, at-first spontaneously felt, experience of aggression is embedded in a space of understanding by linguistic articulation. I agree with Taylor in stressing the connection between language and emotions; of course language can clarify and differentiate our emotions. But it would be useful to say more about the non-linguistic ingredients of emotions. The open questions are: How exactly are linguistic elements interrelated with the non-linguistic quality of an emotion? And don’t we have to distinguish between different kinds of affective phenomena, such as emotions (like shame), moods (like melancholy), and feelings (like pain), in order to get clear about the role of language?
Taylor’s book elucidates the role and scope of human beings’ linguistic capacities on many different levels. The result is a philosophical anthropology that enquires into the essence of the human condition by taking the linguistic skills of humans seriously. I read it as a promising alternative to the often one-sided and historically blind conceptions in the contemporary philosophy of language and mind, where ghostly and worldless creatures populate the philosophical landscape. However, beside the merits of the book, questions concerning the differences between the mediational and the constitutive view of language, considerations about concepts, and a discussion about language and emotions deserve further elaboration.