Vous êtes sur la nouvelle plateforme d’Érudit. Bonne visite! Retour à l’ancien site


Coleridge, Language, and Imagination

  • Nicholas Reid

…plus d’informations

  • Nicholas Reid
    University of Otago

Corps de l’article


The rise of linguistic theory in recent decades has allowed us to see much in Coleridge that goes beyond the traditional concern with imagination, most notably in giving us a broader sense of his vital interest in language. If the Bollingen editions reflect a more traditional, philosophical approach, that approach was in a broad sense recentered on language by Elinor Shaffer and Kathleen Wheeler, and a host of more specific studies of Coleridge's theory of language followed from critics such as Adams, Christensen, Esterhammer, Fulford, Hamilton, and Hodson.  [1] Stephen Prickett's discussion of desynonymy revealed just how sophisticated was Coleridge's grasp of the conventionality of linguistic symbols, and James McKusick has pointed to the deep analogies between the foundational verb substantive and God—the 'is' and 'the infinite I AM'. More recently, David Vallins has documented the depth to which Coleridge sees contradiction as implicit in human language.  [2]

I could not now imagine reading Coleridge without this kind of awareness. What I want to qualify, however, is a tendency to read too much into Coleridge's linguistic interests. I am thinking here of a desire to find in Coleridge a linguistic nominalism or anti-realism of the sort common in theoretical circles until the later 1990s; or the desire to find in his works something akin to Wittgenstein's proposition that 'The limits of my language mean the limits of my world'.  [3] Attractive though this view of Coleridge has been, I can see little evidence for it. The early Coleridge usually announced his enthusiasms loudly: think of his Unitarianism, or of his writings on Hartley, Godwin, Berkeley, and so forth. But apart from a gnomic comment on Tooke, which I discuss later in this essay, the early Coleridge writes comparatively little on language per se.  [4] And the late Coleridge, on which this essay will focus, had views on language that were clearly incompatible with any constitutive approach. The sober truth is that Coleridge was well acquainted with nominalism both in the empirical tradition against which he reacted and in its earlier medieval form, and he did not choose to endorse it.

In this essay I shall argue instead that Coleridge did indeed understand the conventionality of human language, but far from making human language the measure of all things, Coleridge believed that its conventionality was a sign of its limitation. Coleridge may speak of God as 'the infinite I AM', but this does not make God either a linguistic entity (an artifact of language), nor even a real being whose nature is essentially one with that of conventional human language. For Coleridge there is a deeply interesting analogy between God as 'infinite I AM' and the role of the copula in logic and human language, but it is merely an analogy between two levels in the great chain of being, rather than a sign of essential identity.

Indeed, this must be so because human language, for Coleridge, is a product of the understanding, and as such it shares the limitations of that faculty. The understanding plays no part in the existence of Coleridge's God, being, as I shall show, confined to the finite or human sphere, and emerging (with conventional language) only from the delusory efforts of the Apostatic will to find a self-as-object that is distinct from God. Moreover, the conventionality of human language, for Coleridge, reflects our inability to reason in a purely intuitive or immediate fashion; for as limited beings we are constrained to operate through the understanding and to use reified or conventional tokens. Those tokens, however, are indexed to something outside language. While in the Logic Coleridge rejected the empiricism of the ancient view, going back to Aristotle, that words gain their meanings from images, in the Biographia he speaks of the 'voluntary appropriation of fixed symbols to internal acts; to processes and results of imagination' (vol. II, p. 54).  [5] Coleridge indexes words not to 'images', for they are passive empiricist entities, but to the imaging process, a process which is active.

The view of Coleridge that I am putting forward more broadly makes Coleridge akin not to structuralism understood in the radical and anti-realist form of recent decades, but more akin to the views of Susanne Langer. Langer's 'new key' drew on Cassirer, Carnap, and Wittgenstein; she argued that the distinctive feature of twentieth-century philosophy was its insistence on the role of symbols or reified tokens in human thought. But what makes Langer's view unique is her insistence that the 'symbol-mongering' human mind makes use of two kinds, not one kind, of symbol. The first is of course the conventional symbols of language; but the second kind of symbol is the perceptual image, for drawing on work done by L. A. Reid in the 1920s, Langer viewed perceptual images not as passively received empiricist sense data, but as constructed 'presentations'—a term Coleridge also uses.  [6] For Langer these 'presentations' reflect not a single and objectively given way of seeing the world, but rather the human interests of the observer (this does not need to amount to anti-realism as such, for our interest may be in real points of cleavage in the physical or biological world). They are thus 'symbols' rather than passive 'representations'.

In describing sensory percepts as 'symbols' Langer is in danger of reifying them, a point on which Reid had been clearer. Nonetheless, Langer's way of talking does have the virtue of making us see that sense perception is not passive, but is in a way a kind of language, indeed, a more fundamental language which provides the basis for our perception and understanding of the world. This is a way of speaking that Coleridge would endorse, for he too speaks of the formal and sensory as a more fundamental kind of language, being, as he put it in 'Frost at Midnight', 'that eternal language which thy God / Utters'. And this brings us to the core of Coleridge's view of language, for if both kinds are to be called 'language', they must have something in common. For Coleridge, at least from July 1803 and consistently thereafter, that commonality does not lie, as it does for modern theories, in the conventionality of human language. Rather, it lies in the strange fact that 'utterance' is 'outer-ance'—the expression of something which is internal ('idea', 'meaning') in an external form.  [7]

Since Frege and Wittgenstein, the 'psychologistic' view that meaning is linked to image or 'idea' has not been fashionable, though I shall conclude with the suggestion that Coleridge's approach here is the right one.  [8] I shall also suggest (though I must leave the actual argument for another essay) that this reflects Coleridge's belief in a deeper and non-conventional 'language' of form. However, we must start, as always, with the reason/understanding distinction and the Trinity, so as to understand firmly the limited origins of the understanding—and conventional language, its tool.

I. The Understanding's Origin in Dynamic Reason and the Trinity

The Coleridgean understanding (and also Schelling's) derives from the Kantian faculty described in the Critique of Pure Reason. Kant describes, in his preface to that work, his discovery that a priori truths seem to apply within the empirical sphere, but that metaphysical argument invariably terminates in contradiction (that is, the antinomies).  [9] But Kant didn't wish to adopt as broad a scepticism as his predecessor Hume, who had awoken him from his dogmatic slumbers, but who had notoriously reduced even empirical 'cause' to mere habit of association.  [10] Kant wished, more reasonably, to rescue a priori reasoning within the empirical sphere, but was forced to explain how a priori argument can be true of empirical phenomena, while not being true of things-in-themselves (noumena).

Kant solved the problem by adopting what Peirce astutely describes as being effectively a kind of idealism, with respect, that is, to phenomena, though not in any deeper sense.  [11] Kant accepts that a priori principles do describe the phenomenal world, but argues that this can only be possible if the human understanding itself supplies all of the formal (that is, logical, spatial or geometrical, temporal, and mathematical) features of both our sensory perception of, and our thinking about, reality. All thought is thus irremediably subjective, and nothing can be said about the thing-in-itself (the noumenon) as opposed to the thing-as-it-appears (the phenomenon). All transcendent or metaphysical argument is ruled out of court, and the function of philosophy is reduced to supplying a phenomenological (or transcendental) account of the structures of human thought.

Coleridge, however, believed that he possessed what Carlyle sardonically referred to as 'the sublime secret of believing by "the reason" what "the understanding" had been obliged to fling out.'  [12] Coleridge accepted the Kantian description of the uninspirited human mind (reproducing it, for instance, in the Logic), but believed that this describes a faculty that reifies an underlying dynamic reality.  [13]

We can see what Coleridge meant if we look at his view of logic. As I have argued elsewhere, traditional logic takes it as self-evident that the world can be divided into X and non-X (wombats and non-wombats, etc.).  [14] Coleridge, however, observes that if the world is an entity which can be divided in this way, there must be a unity prior to the division. This is in effect what Plato had argued in the Parmenides (as had Spinoza later), though the paradoxical appearance of the argument was one of the grounds for Kant's belief that all metaphysical argument leads inevitably to irresolvable contradiction.  [15] Coleridge, moreover, observes that since the prior unity is prior to the categories of logic it cannot be a thing (since 'things' are defined in terms of the categories of logic) but must be an act. Once its dynamic basis is seen, logical opposition becomes not complete and categorical but an opposition of polarity, in which the polarized elements belong to a single continuum and are the product of an underlying act.  [16] And, more generally, the status of the understanding can be seen by analogy, for all its categories are similarly reified forms of an underlying dynamic act, or what Coleridge refers to as reason.  [17]

The Coleridgean understanding is thus a limited or human faculty that only emerges (along with human or conventional language) at the end of Schelling's dialectic, at the point where human consciousness emerges.  [18] For though Coleridge rejects Schelling's system as an account of ultimate reality, which is why the terms of the understanding cannot be applied to the Trinity, Coleridge is happy to adopt Schelling's system as an account of the human or finite sphere. Schelling had adopted the correspondence theory of truth, a theory that finds in Truth a correspondence between an object and its representation within the knowing subject. Schelling argues that if subject and object are to come into the knowledge relation, they must share a common ontological basis. The steps in the argument can be displayed schematically as follows:

  1. We need a prior ground of subject and object.

  2. Since this is a prior ground it can be neither subject nor object, nor describable in terms of any of the categories of knowledge. It is thus not a thing.

  3. The prior ground must therefore be an act.

  4. As prior to the categories, this act must be without limit—since limitation is a categorical concept.

  5. Knowledge requires a distinct subject and object; and self-consciousness (for Schelling, a putative consciousness of self-as-object) thus requires distinction or limitation. If limitation is to occur, it must imply a second limiting act.

  6. We thus have two infinite activities, identical but opposed in direction.

With the derivation of these two infinite activities, and the limitation of the first by the second, we arrive at Schelling's Absolute, an unconscious, quasi-pantheistic God (or perhaps closer to a neo-Platonic demiurge).  [19] But the limitation of the first act by the second is not the end of the story, for the second act at this point remains unlimited; and a further set of limitations moves dialectically from side to side in an interplay out of which emerge matter and finally human self-consciousness. Nor is this final step a knowledge by a subject of itself as an object, for there are in reality no selves-as-objects in Schelling's universe, and our finite senses of ourselves are in fact the product of an illusory abstraction from the process. Such (transcendental) abstraction requires a faculty of abstraction, and it is for this purpose that the understanding emerges, at the end of the dialectic.

Coleridge, as I have mentioned, did not ultimately accept this argument, for it is based on the possibility of deriving a correspondence of subject and object, something that Schelling admits is ultimately impossible. I have described elsewhere the role of Will in Coleridge's system, and the way in which it grounds a Trinity where the relation is not of subject and object but of distinct persons, each complete in their own right, and the free act by which God grants actuality to the apostatic will. The point I want to make here, however, is that polarity only enters the Coleridgean scheme with the apostasy, and in no way contaminates, or is involved in, the Trinity. As Coleridge says in the 'Opus Maximum', 'the true poles are the apostatic Will and the metathetic or redemptive Spirit and Word' (vol. I, f. 65), not the Father and the Son. And in terms of Schelling's dialectic, this first polarity is the origin of the understanding, for the Satanic principle (the Apostatic Will) is in itself mere chaos, a state without distinction or limit, and is the equivalent of Schelling's first, unlimited act. The act of the Spirit, in giving unity, limit, and indeed form to the first act, is effectively the same as that of Schelling's second, or limiting act. And though both acts are grounded quite differently from Schelling's, they, like Schelling's acts, ground a dialectic from which emerges at the other end (in the moment of reflection) the understanding.

Since conventional language is a product of the understanding, indeed the kind of reified token that that limited faculty is constrained to use, I hope it is by now clear that language is, for Coleridge, not foundational. I stress the point because there has been some confusion over it, particularly in relation to Coleridge's analysis in the Logic of the role of the copula. It has been argued that Coleridge derives the copula (the 'is' which is central to logical judgement, and therefore provides the structure underlying conventional language) from the verb substantive, the sum. This is true, so far as it goes, but we should not read 'derives from' as meaning 'is identical with', and from that draw the conclusion that the sum of the finite mind is essentially linguistic. That conclusion might (or might not) be affirmed of material beings who can form a sophisticated self-concept only through language (the debate is over how much of a sense of self pre-linguistic creatures such as dogs or higher apes can have). But such a conclusion makes little sense in terms of the idealist Coleridge's finite self, which is grounded in a non-linguistic dialectic in which language only features at the terminus—for language is a product of the understanding and arises only after the moment of reflection. The finite self in Schelling's and Coleridge's dialectics may only be able to express its self-identity in language, but that 'expression' (as the etymology of the word implies) is an exteriorising of something which has an origin deeper than language.  [20]

Similarly, it may be that logic must be embodied in language (though again animals display some rudimentary logical thought, without recourse to conventional language). But it only makes sense to say that logic is essentially linguistic if one has a strong theory which says that all of our discriminations in terms of sense and meaning derive from language, and that there is nothing outside language to which these distinctions can be indexed. Structuralism, understood in the strong sense (which is arguably not Saussurean), provides such a theory, for it claims that meaning derives solely from the linguistic structure.  [21] But while Coleridge's logical categories can only be expressed linguistically, the deeper acts of which they are reifications belong in the realm of reason, which precedes the dialectic and conventional language.

If this is true of the finite realm, it follows even more clearly of the relation between conventional language and the 'infinite I AM'. There is a relation, for the infinite I AM is the ground of all being in Coleridge's system, and Coleridge similarly attempts in the Logic to derive all of language from the finite sum, on the grounds that 'the verb substantive ['am', sum] expresses the identity or coinherence of being and act', thus underlying most immediately noun and verb (pp. 16-17). But again we must notice the word 'expresses' in the Coleridge quotation, for 'expresses' does not mean 'is identical with'. Within the Trinity, the Father instantiates himself through the creation of the Son, an 'other' which is essentially his formal embodiment (just as a work of art is a formal embodiment of an idea). This is not linguistic, though as an act which itself grounds Being, the Trinity prefigures and grounds the much later repetition of this in the finite faculty of the understanding. Conventional language doesn't ground being, but in its limited way it expresses or reflects that grounding by joining subject and predicate through the copula.

This needs further explanation—an explanation that takes us to the heart of Coleridge's primary imagination. In the Logic, Coleridge tells us that 'the rose [as a subject] manifests itself, ... renders itself objective ... by its colour and its odour' (p. 82). Similarly, Coleridge claims that substance is the unity of 'accidents', or properties—a claim that has much in common with Leibniz's denial that substance is anything other than a collection of properties, or that sensation involves any contact with an external substance.  [22] In a realist context, it is more usual to think of substance as that which grounds, or holds together, the properties of an object—for how could the rose's colour and odour float around together in the universe if not manifestations of a single underlying, and probably material, substance? But Coleridge is an idealist, for whom no material substances, no physical universe, and ultimately no objects exist: only subjects that are external to the finite self, and are rendered objective when perceived by a finite self. These subjects include other people, but they also include roses and buses and beef-steaks. (It may seem strange to call these latter 'subjects', but in an idealist world there ultimately is nothing outside subjectivity, and so these 'objects' are in fact reflections of the psyche.)  [23]

'Objects' such as roses come to us only through the dialectic. Though they derive from archetypes within the pleroma (the divine ideas), objects appear to us only as a result of the finite self's delusory attempt to perceive itself as object. The finite self is driven through the dialectic by its pursuit of this chimera, a pursuit that comes to a kind of resolution in the moment of reflection, where sensory perception occurs and finite subject and 'object' appear. Physical form and substance are merely the creations of the understanding, the abstracting and reifying faculty that is the condition of the illusory separation of being into subject and object. This means that sensation, in the sensory manifold, is totally particular, since it is prior to the understanding's imposition of form, category, and substance.  [24] And it is only by the affirmation of the object, through the understanding, and as a product of the finite self's attempt at self-perception, that the object gains substance. We perceive not colour and odour separately,  [25] but the rose—an entity that appears in the primary or perceptual imagination as an integrated presentation of these qualities, but an entity that derives its substance only from the mental affirmation which produced it.

That affirmation is itself non-linguistic (for presumably animals also perceive objects in the world), but is reflected in the structure of language where the subject/predicate relation depends upon the unifying role of the copula. We know that it is non-linguistic because it is 'the unity of primary perception', the 'primary mental act' which is 'presupposed ... in all consciousness' (Logic pp. 70, 76; emphasis added), forming the ground of perception but coming before the moment of reflection, the moment when the understanding is formed and from which conventional language emerges.

For to be strictly accurate, we have not yet arrived at the actual perception, at a point in space and time, of the rose itself—since the 'primary mental act' constitutes the second of Schelling's epochs (the first was the creation of the Absolute). It is only through the process of reflection on this primary act that consciousness emerges, a process that necessitates the construction of the understanding; for 'reflection', Coleridge tells us, 'is the understanding itself, a synonym[,] not a predicate' (p. 89). This is Schelling's third epoch; and actual perception only occurs in the moment when a thinking subject consciously perceives an object. The Logic puts it more succinctly: 'Without the repetition or representation of this [primary] act in the understanding [that] completes the consciousness we should be conscious of nothing' (p. 78).

We can conclude, then, that we cannot understand the Logic unless we appreciate the role that Schelling's deduction plays in it, for though the Logic contains no talk of dialectic, its use of the crucial word 'reflection' identifies that context. Once this is understood, we can see that the Logic's talk of the 'primary mental act' looks back to Coleridge's 'primary imagination', and the text effectively provides the argument Coleridge was unable to produce in the Biographia. Far from centering on conventional language, the Logic relegates its place to the end of the dialectic, as a product of the limited faculty of the understanding. It is both interesting and unexpected, as McKusick says, that in one respect language reflects or repeats the structure of the primary affirmation (p. 134). For just as the primary affirmation joins the 'accidents' or properties (colour, scent, etc.) of the rose, so language joins, through the copula, the subject to the predicates which together define it.

It is time to turn to the infinite I AM. In Exodus 3:14, God tells Moses, 'I AM THAT I AM'. Referring to this, Coleridge tells us in the Logic that 'the verb substantive or first form in the science of grammar brings us the highest possible external evidence of its truth' (p. 82). Again, there is a shared structure that we might not have expected. But this is explicitly 'external' evidence, for Coleridge is not saying that 'God' is a linguistic entity, nor even that God's self-realisation through the Logos is a linguistic utterance in the sense that it involves the use of conventional and arbitrary language. Rather, and as we have already seen, the Logos is the formal embodiment of the Father, just as the physical form of a sculpture can embody the artist's conception.

II. A More Specific Examination of Coleridge on Language

I have argued above in rather abstract terms, and wish to turn below to some of Coleridge's more specific comments on language. In general, these show that Coleridge was fully aware of the arbitrary nature of conventional symbols, a view that he first makes explicit in a letter to his wife in May 1799 where he says that 'a Word is but an arb[itrary character]'.  [26] It is also clear that Coleridge saw the language of nature as being based on the more essential or intuitive language of the Logos—that Coleridge's views are in this respect close to those of Langer. But there is a complexity, for in distinguishing the discursive from the intuitive, Coleridge rather surprisingly suggests that they are (to use Raphael's words in Paradise Lost) 'Differing but in degree, of kind the same' (vol. V, p. 488).  [27] The difference between the two kinds of language, we are told in the 'Opus Maximum',

is found not in any diversity of kind in the language of the one or the other but in the transcendent perfection only of that Divine eloquence [Nature].  [28]

vol. II, ff. 41-42

To understand this I think we have to understand that for Coleridge the 'kind-ness' or essence of language lies in something prior to the discursive/intuitive distinction. Where neoplatonic metaphysics tends to think in terms of disembodied 'ideas', Coleridge (as Coburn and Modiano argue) insists upon the Logos as embodied or externalised idea.  [29] And Coleridge, I suggest, was struck by the way in which even conventional words can at times be filled with meaning and appear to transcend their conventionality. At such moments, the 'Opus Maximum' tells us, mind 'outers' itself, for it takes the disembodied idea and locates it in the outward form of a conventional symbol.  [30] This was a view that Coleridge first expressed in July of 1803, with his claim that 'Language & all symbols give outness to Thoughts' (Notebooks vol. I, 1387). Now in human language the symbols are of course merely conventional, for they belong to the understanding, but 'outer-ance', as I have already suggested, turns out for Coleridge to lie at the very heart of being. God instantiates himself not through a disembodied self-knowing, but through the process in which the Father knows himself through the form of the Son. 'Outer-ance', and hence 'language' in Coleridge's special understanding of the word, lies at the heart of the Trinity; and it is in this sense that the languages of nature and of conventional discourse are of the same kind.

In the 'Opus Maximum', Coleridge declares that

if language be, as it is, the offspring and epiphany of the human soul it can not be surprising that the elements of Grammar should supply the most appropriate metaphors for the elementary forms of human reasoning.

vol. I, f. 38

This has a grand ring to it, for Coleridge is again pointing (in the idea of the 'epiphany') to the mystery that meaning can be embodied and externalised in language. But we must not overstate the significance of this, for in its immediate context it is only a passing comment; and in the same manuscript Coleridge defines the soul as the 'synonyme' (sic) of the understanding (vol. III, f. 147)  [31]—a sign rather of its benightedness than a term of privilege. We should also note that where Coleridge calls the elements of grammar 'the most appropriate metaphors', metaphor is for Coleridge a limited mode of the understanding, and its role is explicitly limited here to 'human' reasoning. Nor is Coleridge here suggesting a constitutive role for conventional language: he merely observes that since it takes its origin in the same moment in the understanding where the soul is born, there is a correlation or parallel between its structures and those of the soul. Coleridge goes on to say that for the human mind thought can only be communicated 'by words or language' (vol. I, f. 38). But this does not make thought and language synonymous, for where thought or Idea belong to the immediate or intuitive realm of presentational form, the 'language' of the Logos, words are 'the instruments of communication, ... the only signs that a finite being can have of its own thoughts.'  [32]

Conventional language then is a fundamentally human product, for as Coleridge says, 'All language originates in reflection', and 'reflection', as we have seen, is another name for the emergence of the understanding (Logic p. 234). As Coleridge says in the 'Opus Maximum', 'all language in its very essence is appearance' (vol. II, f. 233)—and the world of appearance, or the phenomenal world, is the world of the understanding. Indeed, as David Vallins shows, amongst Coleridge's many cautionary and critical comments on conventional language is the claim that language inherits the contradictory foundations of the understanding.  [33] Thus, for instance, Coleridge speaks of 'the transcendency of religious Intuitions over Language, which only by balancing of contradictions can represent or rather resuggest them' (Notebooks vol. III, 4183). (We should note that Coleridge here implies that under certain conditions the problem of contradiction can be overcome. Coleridge is not a Derridean manqué.)  [34]

Because words do not in themselves constitute it, Coleridge needs an account of meaning. The familiar though unfashionable correspondence view (that words correspond to and gain their meaning from internal ideas) can be seen as early as February 1805, where Coleridge claims that 'articulate sounds are made by the Reason to represent Forms, in the mind, ['phaenomena']' (Notebooks vol. II, 2445). In the Biographia he tells the same story:

The best part of human language, properly so called, is derived from reflection on the acts of the mind itself. It is formed by a voluntary appropriation of fixed symbols to internal acts, to processes and results of imagination.

vol. II, p. 54

Since Coleridge is an Idealist, the 'results of imagination' encompass (amongst other things) the objects of the natural world, to which the fixed symbols of language are appropriated. Nor should it surprise us that the Biographia's view rests on correspondence, for in that work Coleridge mistakenly adopted the broader correspondence view of truth from Schelling. But if his later manuscripts rejected this view at the level of reason, they nonetheless provide no reason for rejecting it as the basis of conventional language, a product of the understanding. The 'Opus Maximum' thus speaks of the origins of language in 'forms correspondent in kind to the creature' (Nature; vol. III, f. 18);  [35] of 'the sensuous images in which the words we use had their origin' (HM 8195, f. 45); and more broadly of 'forms' of 'the intellect' to which words 'correspond' (vol. II, f. 147). These later views of conventional language, and its dependence on correspondence, are a development of Coleridge's claim in November 1800 that '[a] whole Essay might be written on the Danger of thinking without Images' (Collected Letters vol. I, p. 362).

Readers may object that this is the ancient and impoverished account, dating from Aristotle, which indexed meanings to images, and which neglects the nuance which language is capable of and which cannot be represented in an image. Wittgenstein tells us that a picture of a triangle 'can be seen as a triangular hole, as a solid, as a geometrical drawing; as standing on its base, as hanging from its apex; as a mountain, as a wedge, [etc.]'.  [36] From examples like this, Wittgenstein shows the implausibility of thinking that meanings are indexed to reified pictures in the mind (p. 54). But Coleridge, in the quotation above, is not indexing meanings to reified images: he is indexing them to the imaging process; and how we see an object is capable of all the nuance of language—indeed, all the nuance that Wittgenstein himself points to.  [37] Coleridge's theory is not an empiricist theory, because natural forms are not passively received, as the empiricist account would suggest.

That Coleridge in the 'Theory of Life' describes Nature as 'a more perfect language than that of words—the language of God himself' (p. 486),  [38] or speaks in the 'Opus Maximum' of 'the analogy so beautifully display'd by Bishop Berkly [sic] between the World and a Book' (vol. II, ff. 41-42), will come as no surprise. It is initially a little surprising to find Coleridge saying that the Logos is 'that speech that needeth no translation but the first language [that is, Nature] of which all others are but the imperfect versions.' At first sight this might seem to imply that human conventional language is but an 'imperfect version' of the language of nature and form—to contradict, that is, the Langerian distinction between the two types of symbol.

However, we can again understand more deeply what is going on here if we see the connection Coleridge makes between 'utterance' and 'outer-ance' (vol. I, f. 38). As he says in the 'Opus Maximum', thought can only be communicated 'by words or language' but 'all language is utterance i.e. outer-ance.' This includes the language of nature, for in the Friend he says that sensation, in itself, is, like words in themselves, 'mere empty sounds' significant only as the 'exponents' (outward puttings) of mind (vol. I, p. 440).  [39] For Coleridge, then, language is not characterised necessarily by the mere conventionality of its symbols. Rather, it is the 'outer-ance' of conventional, discursive language which renders it the same in kind as the language of nature; and this similarity extends even to 'that speech that needeth no translation', the Logos (vol. II, p. 42). For that too is a form of 'outer-ance', a relation in which outward form, the forma formata of the Son, embodies the inward principle, forma formans, or the Father.

But if all these are forms of 'outer-ance', there nonetheless remains a distinction between the conventional meanings of discursive symbols, and the essential meaning of the divine language, for the latter is grounded in the Trinity, and there is never any suggestion in Coleridge that its form is arbitrary. Coleridge points to both the similarity and the distinction in Notebook 56,  [40] where he insists that both forms are languages, but that conventional language operates by standing a symbol in opposition to the 'thing intended'. This distinction would make no sense if Coleridge believed that the divine language was conventional:

And tho' the language [of Nature] should not, like Languages of Sound or of conventional shapes, stand in opposition to the Things intended ... Still it is a language.

The Langerian distinction between the two kinds of language is also captured in Coleridge's analysis of those moments in which conventional language seems completely filled with meaning, moments when, as Wallace Stevens put it, 'The words were spoken as if there was no book'.  [41] These are moments when language is enlivened, as Kathleen Wheeler has taught us, by the act of the reader. But Coleridge's analysis here depends on the very strength of his (quite unmodern) belief in correspondence. In the Philosophical Lectures Coleridge speaks of the mystified native poring over the Bible until the missionary translates it into assimilable Idea: then 'the words become transparent and he sees them as though he saw them not' (Philosophical Lectures p. 366). This might be translated as the claim that we think through language rather than in it. The basis of Coleridge's position, for all his wonder in the face of the surprising fact that words can seem to embody meaning, is that words are only means of indexing the real locus of meaning, the presentational form (the language of the Logos) and the Idea it embodies. A conventional language enlivened, I suggest, is a proxy for the presentational forms themselves, and in this sense there is no 'diversity of kind' between the two.

So far I have perhaps tended to downplay the role of conventional language in thinking, for I do not think (and I can see no evidence that Coleridge thought) that we think immediately in conventional language. If we did, the languages of art and form would not speak to us, and Coleridge would not have had any grounds for thinking that the language of the Trinity, like its copy, the language of nature, was a more fundamental language of form. I am not, however, denying that conventional language is an instrument which we, as finite beings, need if we are to think about matters which are at all abstract.

This, I think, was the point of Coleridge's well-known statement in Aids to Reflection that his primary purpose was '[t]o direct the Reader's attention to the value of the Science of Words' (pp. 6-7).  [42] Coleridge adds the following, with a teasing reference to Ezekiel:

Horne Tooke entitled his celebrated work, Epea pteroenta, Winged Words: or Language, not only the Vehicle of Thought but the Wheels. With my convictions and views, for epea I should substitute logoi, that is, Words select and determinate, and for pteroenta zoontes, that is, living Words. The Wheels of the intellect I admit them to be; but such as Ezekiel beheld in 'the visions of God' as he sate among the Captives by the river of Chebar. 'Whithersoever the Spirit was to go, the Wheels went, and thither was their Spirit to go; for the Spirit of the living creature was in the wheels also.'

This again does not amount to a constitutive view of language, though it takes us to the heart of Coleridge's view of words as necessary discursive and conventional symbols—necessary because the human mind is always trapped within the confines of the understanding.

To understand what Coleridge meant in accepting words as the 'wheels of the intellect' (or in thinking in the much earlier letter of words as 'things') we need to see how Coleridge modifies Tooke's terms.  [43] Where Tooke had spoken of words as winged, as a useful but dangerous mode of abbreviation, Coleridge speaks of words as select and determinate, implying both a selector and a standard of reference by which words are determined. The tone in which he says of words, '[t]he wheels of the intellect I admit them to be: but ...' place us on guard against any simple constitutivist reading of the passage, for while words are clearly inseparable from meaning ('Whithersoever the Spirit was to go, the wheels went')  [44] it does not follow that meaning is no more than a matter of words. Coleridge's mention of the Spirit introduces the Trinitarian context, and as I suggested above, to see only the role of words in Coleridge's theory is to neglect the Spirit. Thus for Tooke's 'words' Coleridge substitutes 'living words', for only as enlivened in the mind of the reader do words become the wheels of the intellect. And it is the Spirit which mediates between word and idea, which determines the word. For Coleridge a word is a 'thing' inasmuch as it is one of Langer's conventional symbols—inasmuch as for Coleridge the human mind, caught within the constraints of the understanding, must deal in the counters of the understanding even where its purpose is to transcend the 'fixities and definites' of that faculty.  [45]

We can conclude this examination of Coleridge's view of language, then, by looking at two further passages. The first appears in the Logic, where Coleridge defends metaphysics against the argument that it amounts to 'mere disputes about words' (pp. 119-20).  [46] Coleridge's answer is to aver that this would be an appropriate description if words were no more than 'articulated sound', if they did not have a higher function. He goes on:

For should there be another and higher meaning, namely the meaning itself, should the word differ from the mind only as the breeze differs from the air—should it, I say, mean the truth, contemplated as conveyed and communicated, and the truth itself be a living power [then a dispute about words is important].

This passage says a number of things. It claims that there is a sense in which words are merely 'articulated sound' but another (and characteristically Romantic) sense in which they are the vehicles of meaning. This introduces an extra-verbal constraint, Truth, 'a living power'; and reveals where Coleridge's idea that words 'differ from the mind only as the breeze differ[s] from the air' parts company with the predominant modern view that meaning is purely linguistic. Rather, the passage points to truth 'as conveyed', to words as the conventional symbols necessary for a mind excluded from the immediate intuitive realm of reason and confined within the ambit of the Logic's subject—the understanding.

This Langerian view of human or conventional language granted, we are now in a position to appreciate Coleridge's views on desynonymy. These views can be traced as far back as November 1803, when Coleridge's plans for a work on logic had grown to include 'the whole growth of language', a topic which was to throw light on 'the probable Destiny of the Human Race' (Notebooks vol. I, 1646). But it is in the sixth of the Philosophical Lectures that we find Coleridge explaining desynonymy most clearly, beginning with the claim that:

Words are things; they are the great mighty instruments by which thoughts are excited and by which alone they can be <expressed> in a rememberable form.

Philosophical Lectures p. 257

For this reason Coleridge believes in the importance of desynonymy, for he had argued earlier in the lecture that a desynonymy made by the philosophy of one age becomes the common sense of the next (Philosophical Lectures p. 256). And as he argued in an earlier lecture,

a great idea can [only] be taught gradually, that is by considering it as a germ which cannot appear at any one moment in all its force but will gradually separate the plenitude of its contents into distinct parts.

Philosophical Lectures p. 212

The great idea was Christianity, and the process of its realization requires desynonymy—'the whole process of human intellect' and a process which (presumably) finds its telos in the philosophy of Samuel Taylor Coleridge. As the great idea unwinds, so are the conventional symbols of language indexed more closely, through desynonymy, to the meanings they represent, for it is by words alone that thoughts or ideas 'can be <expressed> in a rememberable form.'  [47] Words are to humans a very necessary instrument, and it is probably helpful to think of them in Saussurean fashion as forming a grid of difference. But the structure of that grid (the meanings it embodies) is not self-generated: it is a product of 'thoughts' or the human act of thinking.

There is another reason for thinking that words alone are not sufficient. For Coleridge, as Wheeler pointed out, knowledge is not interchangeable like tokens on the market.  [48] Words, considered as dead forms, are not capable of delivering meaning to the passive listener. Rather, reading must be inspirited if meaning, an act of mind, is to be drawn from the words. Coleridge makes the point explicit in an appendix to On the Constitution of the Church and State, where he defines the ideal of the transcendental philosopher in the person of a fictional character called Mystes—

one who muses with closed lips, as meditating on Ideas which may indeed be suggested and awakened, but cannot, like the images of sense and the conceptions of the understanding, be adequately expressed by words.  [49]

p. 165

Words do not in themselves embody the intentionality which is the core of meaning; and to reify them (to treat them as intransitive when they are fundamentally transitive entities) is to mistake their nature. It is only when inspirited that the unlettered native is able to read. In the Lay Sermons Coleridge brings together the themes of spirit or grace, readership, and the dead word in speaking of

those inward means of grace, without which the language of the Scriptures, in the most faithful translation and in the purest and plainest English, must nevertheless continue to be a dead language: a sun-dial by moonlight.  [50]

p. 57

Inspiration, we may conclude, equates (albeit under special and constrained circumstances) with what today we have learned to call the act of reading. But, remembering Coleridge's arguments towards the Trinity, we need not interpret the Spirit solely as a metaphor for the act of mind by which the reader interprets conventional symbols. If I am right in identifying language for Coleridge as merely the proxy for correspondent presentational forms, then an inspired reading is one that renders presentational form transitive or alive. The Spirit, we recall, was the unity or communion between Father and Son, or the (relatively) ab intra and ab extra. In describing Coleridgean humanity as caught within the understanding and thus constrained to the use of mediate or symbolic means of communication, I have described humanity as caught within the confines of the ab extra. For interpretation to proceed, it is the ab intra which must emerge, and in Coleridge's metaphysics this can only occur in the terms of the Spirit.  [51]

We may make a further point here. Where Coleridge speaks of conventional language as the necessary means of communication between finite minds I have implied that the real locus of thought lies in presentational form. This is so, but we can strengthen the parallels with Langer's 'symbol-mongering'  [52] mind by recalling that for Coleridge reason and imagination in the human mind are not faculties but powers which must operate within the understanding. As such, they cannot work immediately in the realm of the ab intra, but must generate the ab intra from within the reified symbol, the ab extra. It is in this sense, and only in this sense, that Coleridge claims that meaning is inseparable from symbols; though even here he should refer to the symbols of Nature and Art as well as those of conventional language.


I have argued that Coleridge thinks of the essence of language not as lying in its conventionality, nor in a Sausurrean structure of arbitrary difference. Rather, for Coleridge, the surprising and essential thing about language is its ability to 'outer' that which is inner-meaning. For many readers this will make Coleridge a less interesting theorist, for not only have I placed the imagination back at the core of Coleridge's thinking (with the meaning of words indexed to the imaging process) but I have also indexed meaning to what are essentially 'ideas', entities which Frege banished under the rubric of 'psychologism'. Frege's argument molded twentieth-century philosophy, for it was the basis of a devastating review of Husserl's early Philosophy of Arithmetic (1891), a review that changed Husserl's views, and placed the argument against psychologism at the heart of the Continental tradition.  [53] Frege had a similar influence on the Russell circle, and thus on Anglo-American philosophy, but his greatest influence was on twentieth-century linguistic theory, through the writings of Wittgenstein.

However, I have been careful to speak of the way Coleridge indexes meaning to the imaging process, rather than to 'images' per se, for I do not believe any more than Coleridge does that images, as reified entities, exist. I hope to show elsewhere just how closely Frege's and Wittgenstein's arguments depend upon the notion of just such reified 'images'; and just how weak their arguments are once we move away from such a conception. I also hope to show how recent developments in theory of mind have moved back to talk of consciousness, but with an insistence on its embodiment and thus its evolution out of feeling and sensation.

I shall also talk about the ways in which the artificial intelligence industry is shifting its paradigms for knowledge away from syntactically structured algorithms towards a conception of intelligence as something grounded in the perceptual process. It may be that Frege was right about psychologism. If that is so, all I can do is to insist upon the historical facts, and urge that we avoid an anachronistic co-option of Coleridge as 'our contemporary'. But I believe that the Coleridgean view, shorn admittedly of its idealist and theistic roots, is ripe for revival.

Parties annexes