I caught this insight on the way and quickly seized the rather poor words that were closest to hand to pin it down lest it fly away again. And now it has died of these arid words and shakes and flaps in them – and I hardly know anymore when I look at it how I could ever have felt so happy when I caught this bird.
– From Beyond Good and Evil (1886) by Friedrich Nietzsche
‘What is it about the proposal that strikes me as so disturbing?’ Reading through an article describing a local government measure, I feel opposition rising within me. Normally, forming an opinion about such things would take me some time. But not here. The proposal instantly strikes me as unjust. My reaction is not just intellectual; it is visceral. My emotions are engaged. My imagination is exercised. As I imagine the proposal playing out in practice, the distinctive brand of injustice seems to be jumping out of every word on the page.
I decide to sort out my problem with the proposal in writing, by replying to the colleague who forwarded me the article. ‘It’s unfair!’ Impatiently, I blurt out the kernel of what bothers me about it. But the statement is so general as to be almost empty. ‘Heavy-handed. Quietly authoritarian. Positively harmful.’ More words suggest themselves to me and, after a few false starts, I regain my confidence and press the formulation forward with each sentence. I edit some words, and the correction puts everything in order. Reading over what I wrote, I recognise that, even though there is room for elaboration, at this moment these words accurately capture my position. I have found the words to express my thought.
The gulf between our solitary thoughts and the words that would convey them to others constantly confronts us all. The thoughts we struggle to articulate might be as momentous as a transformative moral epiphany or as ordinary as an insight into a movie or the hurtful behaviour of a friend. They might seem hopeful or alarming, frivolous or serious, lead us to find value in certain things, or worry about others. They might be thoughts that we long had but never articulated or instantaneous insights in which something entirely new and unfamiliar suddenly comes to mind. In many cases, we articulate these thoughts in order to get clear on what they are; we wouldn’t bother making the effort if they were clear to us already.
The experience of getting clear on a thought, with the help of language, has received surprisingly little scrutiny. Philosophers of knowledge influenced by René Descartes have focused almost exclusively on cases in which our knowledge of our thoughts is effortless and instantaneous. For example, turning the key, I might think to myself that the door is shut. No sooner than I have this thought, I know that I think it. While I could be wrong about the door (my lock might be broken), not even a throng of neuroscientists could shake my conviction that I am having this thought. Impressed by the special security of our knowledge of our thoughts in such cases, philosophers have sought to understand it and use it to lay the foundation for all our knowledge. The hard cases, in which we must work to get clear on our opaque thoughts, have gotten far less attention.
These cases were similarly neglected in other fields. Linguists, who have studied the abstract rules of grammar and meaning that allow us to comprehend a boundless range of novel thoughts, have uniformly evaded the question of how we apply such rules to produce utterances. Noam Chomsky, who revolutionised the study of the principles that underlie our grammatical competence, wrote in 1986 that ‘with regard to the far more obscure production aspect … it is difficult to avoid the conclusion that serious problems are touched on here, perhaps impenetrable mysteries for the human mind.’ Those who did dare to investigate the process of turning thought into speech – such as the psycholinguist Willem Levelt in his pioneering Speaking: From Intention to Articulation (1989) – have largely done so by analysing common slips of the tongue (eg, ‘left’ instead of ‘right’, ‘wish’ instead of ‘fish’) in cases where articulation is quick and devoid of any sense of discovery. Without a comparable method to investigate the hard cases, the prospect of studying them didn’t even arise.
And yet, venturing to investigate these cases can illuminate the deeper challenges that we face in articulation, transform our conception of ourselves and our relation to our own thoughts, and help us develop our ideas in other creative pursuits. Familiar as these cases are, they invite some basic questions: what is it for a thought to be clear? What made our initial thought unclear? And how do we make a thought clear, in the relevant sense? These questions engage fundamental issues about the relation between thought and language, and between the unconscious and conscious mind. Our way into them starts with two observations that seem to contradict each other.
We need to chisel away at imprecise formulations, while guarding against words that blur what we think
The first observation is that articulating our thoughts, in the hard cases, is our way of discovering what we are thinking. The philosopher Daniel Dennett in 1991 quoted E M Forster’s quip ‘How do I know what I think until I see what I say?’, affirming that ‘we often do discover what we think … by reflecting on what we find ourselves saying.’ Whether or not I have found a genuine flaw in the government measure from my initial example, I feel that I have gained some insight into what bothered me about it.
The second seemingly contradictory observation is that articulating our thoughts, in the hard cases, is a purposive activity that doesn’t simply consist in producing words mechanically, in a kneejerk way. The words that immediately come out of us when we are struck by our thoughts (eg, ‘How outrageous!’, ‘What a mess!’) might hardly reflect what we think at all. They could come to us as a result of habit, their repetition by other speakers, or just our affinity for the way they sound. The danger of giving in to the mindless flow of such words was highlighted by George Orwell, who in ‘Politics and the English Language’ (1946) warned that the buzzwords that fly in most readily will ‘construct your sentences for you – even think your thoughts for you, to a certain extent – and at need they will perform the important service of partially concealing your meaning even from yourself.’ To succeed in articulation, we need to chisel away at imprecise formulations, while guarding against any words that would blur or change what we think.
The careful selection that we exercise in the process stands in tension with the ignorance that we hope it will remedy. The point of searching for words, in the hard cases, is to clarify what we’re thinking; and the clarity that we’re after seems to consist in the knowledge that we’re thinking some specific thought. At the same time, our choices of words make sense to us, and so it seems that we must make them for a reason. But it is hard to see how we could have a reason to accept or reject any words if we don’t already know which thought we’re trying to express.
Compare: in describing a picture or translating a sentence into another language, we have the picture or sentence clearly in mind and search for the words that would fit it. We can’t select the appropriate words unless we know what the picture depicts or the sentence says. So, if our goal is to express a particular thought, it’s unclear how we could select the appropriate means for achieving it, if we’re ignorant of what we’re thinking. We can’t identify our words as correct without comparing them with the thought, and we can’t compare them with our thought unless we know what thought we’re trying to articulate. Jean-Paul Sartre alludes to this paradox, which we might call the ‘paradox of articulation’, in Being and Nothingness (1943):
This is indeed what linguists and psychologists have perceived … they believed that they discovered a circle in the formulation of speaking, for in order to speak it is necessary to know one’s thought. But how can we know this thought as a reality made explicit and fixed in concepts except precisely by speaking it?
And even if we serendipitously stumble on the right formulation – eg, in the mouth of a friend or on an internet discussion forum – how will we know that it captures what we had in mind?
To try to resolve the paradox, one might point out that language functions not only as a medium for expressing thoughts but also as a means for developing them. The act of expression often exposes gaps and sloppiness in our thinking: ideas, once spoken or written down, can turn out to be less compelling than they first appeared. As soon as we try to articulate these thoughts, our confusion becomes apparent. This common experience could naively tempt one to think that, in all cases where articulation is hard, the formulations that we eventually arrive at add something new to our initial thoughts. Clarifying what we think, according to this view, might not lie in expressing our settled thought but in making up our minds about an issue, by constructing a thought that is more definite and coherent. If our goal is not to produce words that match our thought, there doesn’t appear to be a paradox in accounting for how we manage to recognise the correct words to voice our thoughts.
Things are not so simple, however. While it might, in some cases, dissolve the paradox to view the process of reaching clarity as the construction of thoughts, it is, at best, only half the picture. Our thoughts can be more definite than what we can readily articulate. The mathematician William Thurston, who in 1982 was awarded the Fields medal for his pioneering contributions to geometric topology, wrote that ‘there is sometimes a huge expansion factor in translating from the encoding in my own thinking to something that can be conveyed to someone else.’ Meanwhile, according to the mathematician Nicholas Goodman writing in 1979:
some of the hardest work a mathematician does occurs when he has an idea but is, for the moment, unable to express that idea … Often such ideas first manifest themselves as visual or kinaesthetic images. As the mathematician becomes clearer about them, as they become more formal, he may discover that they manifest considerable internal structure, which is, so to speak, not yet symbolically encoded.
And Nietzsche’s writings are filled with laments about the inadequacy of language to render his most cherished ideas entirely in words. One need not be a mathematical or philosophical visionary to have felt this frustration.
The paradox echoes Socrates’ puzzle: how can we investigate something if we don’t know what it is?
As cognitive science increasingly reveals, our thinking doesn’t run on a single track, like a serial computer, but seems to be organised into a variety of facilities, or modes of thought, that loosely communicate with each other. The jagged nature of the interaction might be responsible for the sense of fissure within the mind, reported by many writers and thinkers. Language is just one mode of thought, with its own characteristic parameters and limitations. Though it uniquely affords us with a distanced perspective on our thoughts, it is only an imperfect instrument for capturing them. There are other modes that can present us with aspects of reality and interface more directly with our emotions but are less amenable to explicit reasoning and articulation. Only an uncooperative (and mean-spirited) interlocutor would regard our difficulties in articulation as a sign that we lack anything meaningful to say.
The spectrum between cases in which we start with definite thoughts and cases in which we construct thoughts in the process of expression spans the full gamut of creative pursuits. On one side, there are painters such as Jackson Pollock and Gerhard Richter, who minimise conscious control over their chromatic abstractions, allowing the medium, and chance, to dictate the result. On the other side, there are those who subordinate even the smallest details of their work to their initial vision. Stephen King compared the process of writing his novels to the excavation of a fossil that was already there when the writing had begun. And Marcel Proust described the musical motifs guiding Vinteuil, the composer in Swann’s Way (1913), as ‘ideas veiled in shadows … perfectly distinct one from another, unequal among themselves in value and in significance’.
The cases in which we use language to construct thoughts might be immune to the paradox of articulation, since they don’t involve discovery of what we were thinking. But that still leaves us with the other side of the spectrum: how do we recognise the formulations of our shadowy thoughts, without knowing what these thoughts are? A closely related paradox arises in other creative domains in which we might begin with relatively complete ideas or ‘fossils’ that guide our work. Without fully knowing the fossil we seek to uncover, we can’t tell whether our product conforms to it. But knowing exactly what we’re seeking would drain the creativity and suspense out of the process of excavation – we might as well outsource the rest of the work to someone else. The paradox echoes Socrates’ ancient puzzle about enquiry in Plato’s Meno: how can we investigate something if we don’t know what it is? And if we do know, what’s the point of investigating it?
Socrates’ solution to his puzzle was to equate enquiry with recollection – a process of resurrecting already acquired but submerged knowledge. Though obviously unsatisfactory as a response to the original problem it was intended to solve, this solution provides a clue to the paradox of articulation. In probing into our thoughts, with the help of language, we don’t start from a cognitive blank. Just as we can discover the full geography of an island by knowing its coordinates and sailing there, we can learn more about a thought by drawing on a certain kind of knowledge of it. The knowledge that we draw on is not the explicit information that one finds in textbooks but a form of implicit knowledge, more akin to first-hand familiarity.
We can get a better handle on what this knowledge is and how it allows us to recognise the correct words by an analogy with simpler forms of recognition. Take our recognition of colours. When we see a particular shade of colour, we might know that it’s the colour between red and white, our favourite colour, the colour of flamingoes and cherry blossoms, and so on. However, the bits of explicit information that we have at our disposal might be less specific than our implicit familiarity with the colour, which we have through our experience of it. Our explicit description of a colour might equally fit a subtly different shade of pink; yet our experience of the colour could be richer than our description and still allow us to tell the two shades apart.
Like our recognition of colours, our recognition of the words that match our thoughts is not based on reasoning from explicit information. In both cases, the click of recognition results from an immediate experience. Just as we can reidentify a colour by relying on a trace of our experience of it, we can recognise a thought in the words that express it by relying on its ‘signature’ – the distinctive way it imprints itself on our experience. Since our recognition is based on an experience rather than on explicit information, our initial (explicit) conception of the thought can be minimal and occasionally misleading. Unlike colours, however, our words can surprise us by uncovering layers of richness and variety in our thought that were hidden from us before.
Such discrepancies between our conception of what we’re looking for and what we end up finding might seem mysterious but are actually commonplace in the case of memory recognition. Think of trying to recall a name, for example, of an actor. When we begin, we might feel that it starts with a ‘T’: ‘It’s Thomas something-or-other…’ But once the right name comes, we realise that it starts with a ‘D’ instead: ‘Aha! It’s Daniel Day-Lewis!’ In many such cases, it would be better to let go of what we expected the name to be. Analogously, trying to mould the formulation to our expectation of what it should be could hinder progress – a lesson that resounds in other creative domains.
I might allow myself to smash a vase in anger, but I don’t decide to do it to optimally express my mental state
Understanding how experience generates a form of implicit knowledge and how this knowledge gives rise to explicit understanding can fundamentally shift how we think of ourselves and our relationship to our own thoughts. We saw that our explicit conception of our thoughts plays a limited role in the hard cases; our immediate experiences of the thoughts seem to lead the way. But if the explicit reasons, purposes and plans with which we identify don’t guide our word selection, we might worry that the mechanism I have described leaves no room for us. And yet, there is still a crucial way in which we’re caught up in this process.
Consider an analogous case: emotional expression. Unlike involuntary bodily changes that are part of an emotion (eg, blushing, perspiring, trembling), our expressions of emotions (eg, jumping with joy, ruffling a child’s hair in affection) often seem to be things we do intentionally, even if not in response to explicit antecedent plans or reasons. In her account of emotional expression, the philosopher Rosalind Hursthouse in 1991 argues that many actions expressive of emotions can’t be explained in terms of reasons at all: I might allow myself to smash a vase in the grip of anger, but I don’t deliberate and decide to do it for the reason that it would optimally express my mental state.
Similarly, articulating my objection to the government measure in my initial example doesn’t involve explicit reasoning concerning what its formulation must be. My conscious contribution to the process is at first one of trial and error, in which I attempt to spark a spontaneous process of expression and then allow it to unfold. Just as the process of expressing an emotion is directly controlled by the emotion, the process I attempt to bring about is directly controlled by my thought.
Nevertheless, far from running on autopilot, articulating a thought takes sensitivity, flexibility, attention and care. Articulating my objection to the government measure is manifestly something that I do rather than something compulsory that overtakes me. Although the process is controlled by the thought, it is simultaneously controlled by me. There lies an intriguing feature of our involvement in articulation. Once the process is underway, I can become absorbed in it and experience myself as intentionally carrying it out. The words that I produce are deliberate not in the sense of being deliberately selected but in the sense of being unimpeded by internal censorship or constraint. This understanding of articulation provides a way out of the paradox by showing how we can not only recognise but, also, actively produce the words that express our thoughts without drawing on any explicit knowledge of what we’re thinking.
Whether or not one accepts this solution, the paradox gives us a tool to investigate the process of articulation in a systematic way. The investigation stands at the crossroads of two core interests – the nature of knowledge, and the nature of agency and the self. Arriving at an understanding of this process is not just an intellectual exercise but a practical pursuit of trying to uncover the foundation of how we come to know the world and ourselves. If we’re ignorant of the underpinnings of our own responses, we’re controlled by them; getting clear on the thoughts that push us to these responses can lead to liberation. The theoretical struggle of turning our murky familiarity with this process into explicit understanding is itself within the scope of the present investigation; these words are a product of this same process.