Methods and Tools in Language

optional Theme - Language

Does the language you speak shape what you can know?

THE PROVOCATION

How does language produce knowledge - and what escapes it.

"We know more than we can tell."

Michael Polanyi, The Tacit Dimension (1966)

Listen to the short recording below of me reading a poem by my favourite poet, Dylan Thomas. Do not try to make 'sense' of it or even read the words. Just listen, or as the author famously said, 'love the words.'

Once it was the colour of saying - Dylan Thomas

The poem is called "Once it was the colour of saying." It was published in 1939, when Dylan Thomas was twenty-four - already his third book of poems. In fact most of his published poems were either written or derived from a series of notebooks he kept in his bedroom between the ages of fifteen and nineteen. You are approximately the same age he was when he was building the craft the poem reflects on. The poem looks back to those notebooks - the exercise books he filled in Swansea in Wales, feeling his way by ear through thousands of lines. Once it was the colour of saying, not the meaning of the words, that mattered.

For Dylan Thomas his poetic knowledge was already working at fifteen. The question this page is built around is: what kind of knowledge is that? The harder question is what Thomas knew, in the sounds he was choosing and the rhythms he was building, that produced what you just heard. Can you state it? Could he? If you asked him to teach you how to do it, what would he say?

These are questions about method. This page is about what language can do as a tool for producing knowledge, and about what it cannot do - about the knowledge that lives in the doing rather than in the saying.

Big idea 1 - The body knows before language does.

Michael Polanyi was a physical chemist who became a philosopher of knowledge. His central observation was that a very large portion of what we know cannot be fully stated in language. He called this tacit knowledge: the knowledge carried in practice, skill, and perception rather than in explicit spoken or written rules or propositions. The classic examples are physical. A cyclist knows how to balance without being able to explain the physics of what they are doing. A surgeon knows where to cut from years of practice in ways that cannot be fully transmitted in a textbook. In each case the knowledge is real, demonstrable, and transmissible - but it is transmitted by doing alongside an experienced expert, not by following written instructions. Polanyi called this the apprenticeship model of knowledge.

The French philosopher Maurice Merleau-Ponty, writing in Phenomenology of Perception (1945), provided the explanation for why skilled knowledge resists language. The body itself is a knowing subject, with its own way of orienting toward and engaging with the world - what Merleau-Ponty called motor intentionality - that precedes and operates independently of conscious verbal thought. A skilled typist's fingers find the keys without waiting for the mind to direct each movement. A musician's hands know the phrase before the player has consciously formulated it. Kahneman's framework maps onto this directly. System 2 - slow, deliberate, verbal - is how skills are initially acquired. Mastery consists in that knowledge migrating to System 1: fast, automatic, and no longer dependent on language to operate. The expert typist and the experienced surgeon have moved their knowledge somewhere language cannot easily follow - the expertise is operating at full capacity, but below the threshold of verbal articulation. This is precisely why Polanyi's apprenticeship model is the only reliable way to transmit it.

Language offers its own clearest example. You speak your native language with near-perfect grammatical accuracy. You could not write down the rules you are following. You could not, without years of linguistic study, explain why "the big red ball" is natural and "the red big ball" is not, though you know immediately that one is wrong. Noam Chomsky called this the distinction between linguistic competence (the knowledge you have) and linguistic performance (what you can produce and judge). The competence massively exceeds the ability to describe it.

Merleau-Ponty's most striking case is the phantom limb. A patient who has lost an arm continues to experience sensation and voluntary movement in the absent limb. The body's schema - its map of itself - does not update immediately when the physical limb is gone. The patient knows the arm is there in a way that precedes any propositional knowledge of its absence. Body-knowledge can operate independently of what the mind explicitly knows, and can even contradict it.

Can you still feel a limb that's gone?
TED-Ed

The vast majority of people who’ve lost a limb can still feel it, not as a memory or vague shape, but in complete lifelike detail. They can flex their phantom fingers and sometimes even feel the chafe of a watch band or the throb of an ingrown toenail. What causes these phantom limb sensations?

Joshua W. Pate explains how the brain reacts to a missing limb. Lesson by Joshua W. Pate, directed by Kozmonot Animation Studio.

Dylan Thomas at eighteen had an extraordinary tacit, embodied knowledge of sound - of what happened when certain consonants clustered, when a line turned on a particular vowel, when rhythm broke against meaning. He could exercise this knowledge with precision. He could not, and did not, fully explain it. His craft notebooks show a poet feeling his way by ear rather than following a stated method. The knowledge was in the body - in the ear and in the hand - before it was in any proposition about what good poetry requires.

Polanyi's answer to whether tacit knowledge can be transmitted was yes - but only indirectly, through imitation and guided practice alongside someone who already has it. You cannot learn to ride a bicycle by reading an explanation of balance. You cannot learn to write like Thomas by studying his grammar. The method of transmission is apprenticeship. Merleau-Ponty's framework explains why: the knowledge lives in the body, and the body can only acquire it by doing. What does this imply for schools, which are built around explicit, statable, and assessable knowledge?

Big idea 2 - Speech and writing are not the same cognitive act.

When you heard the Dylan Thomas poem, something different happened from reading the poem on a page - and the difference is neurological, not a matter of preference. Speech is ancient and natural. The human brain has innate architecture for producing and understanding spoken language - the neural structures that underpin it are present from birth and develop without formal instruction. Every typically developing child acquires the spoken language of their community without being taught. Writing is approximately five thousand years old, a recent invention in evolutionary terms, and the brain has no innate module for it. Instead, as the cognitive neuroscientist Maryanne Wolf has shown in Proust and the Squid (2007), the reading brain repurposes circuits built for object recognition, spoken language, and memory, rewiring them into a new network. Every child who learns to read is performing an act of neural engineering that took civilisation thousands of years to develop.

This matters for knowledge because the two acts produce different cognitive experiences. Spoken language carries what linguists call prosody: the pitch, stress, rhythm, pace, and breath of a particular voice. These carry meaning - they are what the poem is made of. When I read "the colour of saying," the long vowels and the weight on certain syllables carry meaning that a flat reading would lose entirely. The phonetician's tools - the International Phonetic Alphabet, the spectrogram - can transcribe what the voice does, but the meaning the voice carries exceeds what any transcription can recover. Your teacher reading aloud to students is doing something categorically different from playing a recording, however technically superior the recording might be. The voice in the room carries relationship and presence in ways that a file cannot.

In Plato's Phaedrus, Socrates relays the myth of Theuth, the Egyptian god who invented writing, and Thamus, the king who rejected it. Writing, Thamus argues, will weaken memory - people will rely on external marks rather than internal understanding - and produce the appearance of wisdom without the substance. A text cannot respond to questions. It cannot adapt to its reader. Socrates argued a text is a dead thing that mimics knowledge.

Walter Ong and Jack Goody spent their careers demonstrating that Thamus had it exactly backwards. Writing restructures the consciousness that produces thought, rather than recording thought that already exists. Certain kinds of thought are only possible because writing exists. The logical syllogism requires holding multiple steps in sequence and checking them for validity - oral argument cannot do this reliably. The scientific table creates relationships between items by the act of listing them. The philosophical definition can be revised until it holds. Goody called the table a cognitive instrument - something that produces thought rather than displays it. It is why you write notes in your lessons and why those notes work best in your own words. There is also an irony Socrates could not have anticipated: his argument against writing is preserved only because Plato wrote it down.

Walter Ong on Orality and Literacy

Ong's central claim in Orality and Literacy (1982) is that literate and oral minds work differently - not in intelligence, but in the kinds of thought they can sustain. Abstract categorisation, formal logic, the ability to examine a statement independently of the speaker who made it: these are products of literacy, not capacities that existed before writing and were then recorded. Wolf's reading brain is the neuroscience of Ong's argument: what writing produces is not just stored knowledge but a differently structured mind.

Maryanne Wolf's second book, Reader, Come Home (2018), extends her argument into the present. She shows that deep reading - the slow, inferential, empathetic engagement that sustained literary attention produces - builds cognitive capacities that digital skim-reading does not. It would be more effective if you were reading this on paper rather than on a screen. The capacity for inference and the ability to inhabit another perspective are products of a particular kind of reading practice, not of reading as such. The medium shapes what the reader's brain does with the content, and therefore shapes what the reader comes to know.

Wolf's argument about the permanence of reading rewiring has a striking demonstration. Try this test before reading any further. The Stroop Test provides a clear window into the struggle between the two distinct cognitive processes. The task requires you to identify the ink colour, which engages System 2. This is your "Executive Brain," characterized by slower, more deliberate, and logical thinking.

To succeed, System 2 must manually override the impulse of System 1 to read the word, a phenomenon known as "interference".

As the 30-second timer progresses, the task becomes more difficult due to cognitive fatigue. While System 1 is energetically "cheap" to run, System 2 is highly "expensive" and tires quickly.

As your prefrontal cortex depletes its metabolic resources, your ability to inhibit the automatic reading response weakens. Eventually, System 1 begins to win the battle, leading to the "lag" or mistakes you experience as your brain reverts to its easiest, most practiced path.

The Stroop test returns us to Polanyi and Merleau-Ponty. Your reading competence is so deeply embedded that you cannot access it reflectively and ask it not to work. That is tacit, embodied knowledge - not as a philosophical abstraction but as an experience you can feel in the half-second delay between seeing the word and naming the colour. The knowledge is in you, in the body's trained automatic responses, operating without your instruction.

Big idea 3 - Translation is interpretation, not transcription.

If tacit knowledge resists articulation because it lives in the body, and if the medium in which language is received shapes what the reader comes to know, then translation faces a compound problem. It must navigate between languages, between embodied ways of knowing, between voices, between cognitive acts that are genuinely different. Consider what a translator of Dylan Thomas must preserve. The words - yes. The syntax, approximately. The imagery, with effort. The rhythm, partially. But "the colour of saying" in Welsh, or French, or Japanese? The colour of that specific saying - the knowledge encoded in the sound pattern of that particular line in that particular language - cannot cross. The translator must choose what to sacrifice and what to attempt to save. There is no method that makes all of it transferable, because the knowledge that matters most is precisely the knowledge most deeply embedded in the specific sounds and structures of the original. I like to read Dylan Thomas in French. I enjoy it because I know the poems so well in English. Reading fascinates through the translator's choice of words, which both illuminates the process of translation and the original poem.

The American philosopher Willard Van Orman Quine made this argument at its most radical. In his thought experiment of "radical translation" - attempting to translate a language with no prior knowledge of it, from scratch - Quine argued that there is no single uniquely correct translation. The same evidence (the speaker points at a rabbit and says "gavagai") is compatible with multiple translations ("rabbit," "undetached rabbit part," "rabbit-hood manifested here"). Translation is always underdetermined by the evidence. The translator is always making an interpretive choice - a series of decisions about what matters most, with no neutral method to arbitrate between them.

Harry Potter and the translator's nightmare
Vox

Translating the Harry Potter books written by J.K. Rowling, in over 60 languages around the world, was not for the faint of heart or vocabulary. Translators didn't have advanced copies of the books to get a headstart and these books could take months to adapt from English. They also had to be clever in their solutions because the books are filled with wordplays, invented words, puns, British culture references, riddles, and more.

Translation is a method with acknowledged limits - a form of knowledge production that is always partial and always involves the translator's judgment. The best translators know this. Jorge Luis Borges, who translated Whitman, Faulkner, and Woolf into Spanish, wrote about translation as a form of reading so deep it produces a new work. Something is always lost. Something is sometimes gained. The question is whether the loss is acknowledged. For multilingual students, this is not abstract. When you translate a joke, an idiom, a feeling from one of your languages into another, you perform Quine's problem in real time.

Consider the English sentence attributed to Groucho Marx: "Time flies like an arrow; fruit flies like a banana." In English it works because "flies" operates simultaneously as a verb and a noun, and "like" as a preposition and a verb - the sentence appears to be a philosophical observation and turns out to be a statement about insects. In French, German, or Spanish the double meaning does not exist. Any translation produces one reading and destroys the other. The joke is untranslatable - you know from the inside that the sentence you arrive at is an approximation, that something in the original resists the crossing.

What you are feeling in that moment is the boundary of what language, as a method, can do. Merleau-Ponty's framework explains precisely what that something is: the embodied knowledge and felt sense that a word or phrase carries in the language of its origin - and that cannot be loaded into a different linguistic body.

Bringing it together

The limits of the method are part of the method.

The three big ideas on this page share a common thread: language as a method has a fundamental and irreducible limit. Merleau-Ponty and Polanyi show that a great deal of knowledge lives in the body, in trained perception and motor intentionality, in the place where the knower's expertise works before language can reach it. This is the nature of a particular kind of knowledge - one that can only be acquired through practice and transmitted through apprenticeship.

Wolf shows that the medium shapes the knowledge it carries: reading and listening are different cognitive acts, and what each produces is genuinely different. The reading brain and the listening brain are different instruments encountering the same words - and the difference in what they produce is real, not incidental. Translation shows what happens when knowledge encoded in one linguistic and embodied system meets the boundary of another: something always fails to cross, and what fails is usually the embodied knowledge, the felt sense that Merleau-Ponty and Polanyi identify as the deepest layer of what we know.

The limits Wolf, Merleau-Ponty, Polanyi, and Quine identify are constitutive of language as a method - the conditions under which it operates, not signs that something has gone wrong. A method that acknowledged no limits would be claiming more than any method can deliver. The value of these thinkers for TOK is precisely that they map the territory honestly: they show what language can do with great precision, and therefore also where it stops. Thomas knew in his hands and his ears what he could not state in any proposition about poetry. He was working at the edge of what language can do. The poem published in 1939 is him at twenty-four looking back at what he had been doing. Whether what the notebooks contain is knowledge is one of the questions this page leaves you with.

Next: Optional Theme 1: Language - Ethics.

Questions, assessments, films and other stuff.

Questions to think about

Polanyi claims we know more than we can tell. Merleau-Ponty argues this is because knowledge lives in the body, not only in propositions. Do these two claims support each other - or is there a version of tacit knowledge that does not require an embodied explanation?
You speak your native language grammatically without knowing the rules. Does this mean your most expert use of language is also your least reflective? What does this suggest about the methods of linguistic self-knowledge?
Merleau-Ponty's phantom limb case shows a body that knows something its owner does not consciously know. If the body can hold knowledge independently of the mind, what does this imply for how we identify the "knower" in a knowledge claim?
Maryanne Wolf argues that deep reading builds cognitive capacities that digital skim-reading does not. Is this an empirical claim about the brain, or an evaluative claim about what counts as knowledge? Can it be both?
The poem on the page and the poem in the voice contain the same words. Are they the same object of knowledge? If not, which is the poem?
Quine argued that translation is always underdetermined - there is no single correct answer. Does this mean all translations are equally valid, or that some judgments are better than others even when no method can settle the question?
Thomas's notebook poems were written when he was between fifteen and nineteen. The tacit knowledge in them was developed young, through compulsive practice. Is there knowledge that can only be acquired at a certain age - and if so, what does this imply for education?
When you translate an expression from one of your languages into another and feel that something has been lost, what exactly is it? Can you name it? And if you cannot name it, does Merleau-Ponty's account of embodied knowledge - or Polanyi's tacit dimension - explain why?

Exhibition connections
See more exhibition ideas and previous student work here

It is never too early to start thinking about your TOK Exhibition. The ideas in this lesson connect strongly to four of the 35 prompts. Start noticing objects in the world around you that speak to these questions.

Prompt #23: How important are material tools in the production or acquisition of knowledge?

Suggested object: a book, a printed poem, a vinyl record, an audiobook, a screen showing a digital text - or two versions of the same text in different media. Wolf's argument is that the material tool shapes the knowledge it delivers: the reading brain built by sustained engagement with printed text is different from the brain shaped by digital reading. The exhibition argument asks whether the object through which knowledge is delivered changes the knowledge itself - and whether this matters for what we count as knowing. Merleau-Ponty deepens the argument: the tool shapes what the body learns to do, and that bodily learning is itself a form of knowledge.

Prompt #10: What challenges are raised by the dissemination and/or communication of knowledge?

Suggested object: a translated poem or literary text alongside its original - or a bilingual edition where both versions are visible. The argument draws on Quine: translation is interpretation, and what crosses the linguistic boundary is always partial. The exhibition object should be chosen so that the student can point to something specific that resists translation - a sound pattern, a cultural resonance, an untranslatable phrase - and use it to ground the TOK argument in something concrete. Merleau-Ponty adds the explanation: what resists crossing is the embodied, felt knowledge that the original language carries in its body.

Prompt #32: What makes a good explanation?

Suggested object: a textbook explanation of a skill, alongside something that demonstrates the skill being performed - a score and a recording, a recipe and the dish, a grammar rule and a piece of writing that follows it perfectly without following any conscious rule. Polanyi's point is that the best practitioners cannot fully explain what they do. Merleau-Ponty's point is that the most important knowledge lives in the body and in the doing, rather than in any verbal account of it. The exhibition argument asks whether a good explanation is one that enables performance - in which case apprenticeship alongside an expert may be more powerful than any verbal account - or one that produces understanding, which may require a different method entirely.

Prompt #9: Are some types of knowledge less open to interpretation than others?

Suggested object: a scientific paper and a poem on the same subject - or a medical diagnosis and a patient's account of the same illness. The contrast between explicit propositional knowledge (the kind that can be stated, checked, and falsified) and tacit or embodied knowledge (the kind that is felt and performed) raises the question of whether interpretive openness is a weakness or a feature. Some knowledge is more open to interpretation because it is more complex - interpretive openness is a mark of complexity, not of unreliability.

Feature films
For more see my 10 films for the TOK journey page.

🎬 WATCH — Arrival (2016)

Denis Villeneuve

Arrival works differently on this Methods page than on the Perspectives page. Where Perspectives uses the film to explore Sapir-Whorf, Methods uses the translation methodology: the systematic, falsifiable, evidence-based process by which Louise Banks decodes the heptapod language. Watch for the moments when the method breaks - when the translation is underdetermined and she must make an interpretive leap. These are Quine's problem made visible. The film also illustrates Merleau-Ponty's point: Louise's knowledge of the heptapod language changes not just her thinking but her perception - the body learns a new way of engaging with time. My students can watch the film here.

🎬 WATCH (or rather listen) - Do not go gentle into that good night (various)

There is no known version of Dylan Thomas reading 'Once it was the colour of saying', which is why you had to put up with my amateur version at the top of the page. But two years before he died at the age of 39 he did record his most famous poem 'Do not go gentle into that good night'. All the great Welsh actors, all shaped by the same oral and choral tradition, have each recorded it since.

Listen to all four in sequence. Then consider:

What does each speaker do with the villanelle's repetition? The two refrains return six times. Are they the same line each time, or does something accumulate?
Burton's version is probably the most famous. Is fame a form of interpretation? Does hearing Burton first change what you hear in the others?
Thomas's own reading is rougher, less polished than Burton's. Does that make it more or less authoritative? What would it mean for a poet's own voice to be the wrong reading of his poem?
This page has argued that translation is interpretation, not transcription. Does the same apply to performance? Is Hopkins performing the same poem as Sheen?
Merleau-Ponty argues that the body knows things the mind cannot articulate. Where in these performances do you hear something that you could not describe in words?