Can Words Capture Truth?

The Kena Upanishad and Wittgenstein arrived at the same boundary, 2,500 years apart.

Kena Upaniṣad vs Wittgenstein. Language can point toward ultimate reality but cannot contain it. The Kena drew the same boundary Wittgenstein would draw 2,500 years later, then walked one step further than he ever did.

The Manuscript in the Pack

November 1918. The Italian front. An Austrian artillery officer named Ludwig Wittgenstein has been carrying a manuscript through trench warfare for most of the war. It travels in his rucksack alongside his ammunition and his copy of Tolstoy's Gospel in Brief. He has rewritten it on whatever paper he could find, in pencil and in ink, between artillery fire and night marches. It is short, about seventy-five pages, organized as a series of numbered propositions from one to seven. He thinks it solves every fundamental problem in philosophy.

Wittgenstein writing in a stone POW barracks at Cassino in 1918

When the Italians capture him on the third of November and march him to a prison camp at Cassino, he keeps the manuscript with him. He finishes it inside the camp. He writes to Bertrand Russell from prison and asks him to find a publisher. The book will appear three years later under the title Tractatus Logico-Philosophicus. Its final sentence, the most famous sentence in twentieth-century philosophy, is seven words long.

'Whereof one cannot speak, thereof one must be silent.'

Wittgenstein had been working with formal logic, alone, in trenches, to draw a precise line between what language could do and what it could not. He arrived at the line and stopped at it. He spent the rest of his life walking around it, looking for what was on the other side, and never quite finding a way through. He was the first European philosopher in centuries to take the limit of language seriously as a discovery, not as a problem to be solved by sharper definitions.

Twenty-five centuries earlier, on the other side of Eurasia, an unknown rishi had stood at exactly the same line. The text he composed is called the Kena Upaniṣad. It is one of the shortest of the principal Upaniṣads, four small khaṇḍas, the first two in verse and the last two in prose. It takes its name from its opening word, kena, 'by whom?'. The Kena rishi did not stop at the line. He told a story about what was on the other side, and he handed his student a method for getting there.

This lesson is about the line and the door. The Kena and the Tractatus saw the same boundary. Only one of them knew where to stand when he got to it.

The Kena's Answer

The Kena's opening question is uncomfortable in its directness. Keneṣitaṃ patati preṣitaṃ manaḥ? 'By whom directed does the mind go to its objects? By whose will do people utter the speech they utter?' The rishi is not asking for neuroscience. He is asking who stands behind the mind, who stands behind the speech, who stands behind the very capacity to ask the question.

The answer comes in one of the most precise sentences in Sanskrit. Yad vācānabhyuditaṃ yena vāg abhyudyate. 'That which is not expressed by speech, but by which speech is expressed, know that alone to be Brahman.' The same formula is then applied to the mind: that which is not thought by thought, but by which thought thinks. To the eye: that which is not seen by the eye, but by which the eye sees.

Notice the structure. The Kena is not saying Brahman is a difficult object the mind struggles to reach. It is saying Brahman is not an object at all. It is the condition for the possibility of the mind reaching anything. To try to know it the way you know a tree is a category error. The mind cannot catch it for the same reason a flashlight cannot illuminate its own source of light. The light is already the medium of every illumination.

Then comes the verse that most directly addresses language. Yasyāmataṃ tasya mataṃ mataṃ yasya na veda saḥ. 'To him by whom It is not thought, It is thought. He who thinks It, knows It not.' If you think you have grasped Brahman, the very fact that you think so is proof that you have not. What you grasped was a mental representation. The real Brahman is what was doing the grasping, and it can never become the thing grasped.

The next verse completes the move. Pratibodha-viditaṃ matam. 'It is known through the awakening.' The word pratibodha is crucial. It is not 'understanding'. It is immediate recognition, the kind of direct awareness that happens when you suddenly see that you are awake. The ultimate reality is not reached through inference. It is reached through a recognition that is not itself a piece of content the mind holds.

And then the Kena does something Wittgenstein never did. It tells a story.

Yaksha holding a grass blade before a humbled Indra in a mountain meadow

In the third khaṇḍa, the gods have just won a great victory over the demons and are becoming proud. A mysterious yakṣa appears before them. They cannot recognize it. They send Agni to find out what it is. The yakṣa places a blade of grass in front of Agni and says, 'Burn this.' Agni, whose very nature is to burn everything, tries with all his power. He cannot burn the blade. He returns humbled. They send Vāyu the wind god, and the yakṣa says, 'Blow this away.' Vāyu cannot move the blade. He too returns humbled. Finally Indra himself goes. The yakṣa disappears. In its place Indra sees a radiant woman, Umā, daughter of the Himālaya, who tells him what the yakṣa was. The victory of the gods had been the victory of Brahman working through them. They had mistaken the action for their own.

The story is not decoration. The gods hear no argument. They undergo an encounter. They learn by being unable to do what they thought they could do. The Kena is showing, in narrative form, that the answer to its opening question has to be delivered by an act of recognition, not by a statement of content. Words point toward Brahman but cannot hold it. The best words can do is drive you to the edge of what words can do, and hand you off to a recognition that is itself not a word.

The Western Echo

The Tractatus reads like a mathematical treatise, but what it is trying to do is draw a line, from the inside, between what language can say and what language cannot.

Wittgenstein's argument, stripped down, is that a proposition is meaningful only if it pictures a possible state of affairs in the world. 'The cat is on the mat' works because it corresponds to a configuration of objects that either does or does not obtain. The logical structure of language mirrors the logical structure of facts. Anything that cannot be set up as such a picture is not strictly meaningful. It is nonsense, in the technical sense of 'not sense'.

And here is where the Tractatus becomes interesting for us. Wittgenstein realized that this account left a whole category of things on the unsayable side of the line: ethics, aesthetics, the meaning of life, the existence of God, and the picture theory itself. The framework that lets him draw the line cannot itself be expressed as a picture of a fact. At the end of the book he admits this directly. His own propositions, he says, are 'elucidatory' and must be thrown away once the reader has climbed them, like a ladder kicked away after the climb.

The final proposition is seven words. Whereof one cannot speak, thereof one must be silent. For fifty years readers took this to mean that ethics and religion were meaningless and should be abandoned. The logical positivists of the Vienna Circle read Wittgenstein this way. They were wrong. In a 1919 letter to the publisher Ludwig von Ficker, Wittgenstein said the real point of the Tractatus was ethical and the important part of the book was the part he had not written, which was everything above the line of the sayable. The silence was not dismissal. It was reverence.

Thirty years later, back at Cambridge, Wittgenstein took apart his own first book. In Philosophical Investigations, published after his death in 1953, he abandoned the picture theory almost entirely. Meaning, he decided, is not a sentence picturing a fact. It is how the sentence is used inside a form of life, a 'language game'. He did not retract the line. He complicated it. The final silence remained his answer to what lies beyond.

Give him his full credit. He arrived at the boundary alone, with no teacher, no tradition, using pure logical analysis. He was working inside a Western tradition that, since Plato, had more or less taken the adequacy of language to reality for granted. He showed, in his own methods, that language has an inside-out limit that cannot be argued away. The Kena rishi would have recognized him immediately.

The Gap

Wittgenstein and the Kena agree on the boundary. They disagree on what to do with it.

Wittgenstein's final instruction is silence. The silence is respectful, not dismissive. It is the only honest response to what lies beyond the sayable. But it leaves the seeker facing a blank wall with no indication of what to do next. Wittgenstein himself was tortured by what this silence meant. He gave away his inherited fortune, worked as a village schoolteacher, considered becoming a monk. He lived the silence as an ethical posture, but his philosophical system gave him no resources for what the silence was hiding.

The Kena's instruction is different. Yes, ultimate reality cannot be captured by words. But the response is not silence as an endpoint. The response is pratibodha, immediate recognition. The text does not leave the student at the wall. It says, in effect: the wall is the wall of conceptual thought, and behind the wall is the very awareness that is currently failing to get through the wall. You do not need to break through. You need to recognize that the one who is failing is already on the other side. The boundary is not between the knower and an unknown object. It is between the knower as an object of thought and the knower as the ground of thinking. The second is closer than your own face.

The Upaniṣad makes three moves the Tractatus does not. First, it distinguishes two senses of 'knowing'. Ordinary knowing, in which you stand in front of an object and frame a proposition, is the kind language serves. Pratibodha-knowing is a different mode entirely. Wittgenstein had no vocabulary for this second mode. His framework was propositional all the way down. When propositions gave out, his account gave out.

Second, the Kena gives the student a positive practice. The yakṣa story is one. Kena 2.3 is another: notice that the one who thinks 'I know Brahman' has just proven, by that thought, that they do not. The Upaniṣad gives the student something to do with the boundary. Wittgenstein gives only the boundary.

Third, and most important, the Kena is backed by a living tradition of practice. The verse sits inside a pedagogical lineage in which teachers guide students through exactly the recognition it is pointing at. The silence at the end of the Tractatus is a personal silence in the mind of a reader holding a book. The silence at the end of Kena 2.3 is a classroom in which a teacher watches the student arrive. A boundary without a practice is just a boundary. A boundary with a practice is a door.

The gap is not a gap of honesty. Wittgenstein was extraordinarily honest. It is a gap of what each thinker had available. He had formal logic and personal conviction. The Kena rishi had a living tradition, a two-layer theory of knowing, and a narrative pedagogy. Both hit the same wall. One of them was standing next to a door.

Why It Matters Today

This is not abstract. It is the defining intellectual problem of the age of large language models.

Billions of people now interact daily with software that generates fluent text about anything. The systems have no understanding of what they say, in the sense that matters to the Kena or to Wittgenstein. They have statistical patterns over tokens. They produce confident sentences about love, death, God, meaning, physics, and breakfast, at industrial scale. The Kena's verse applies in a new way: yasyāmataṃ tasya mataṃ. The thing that produced the paragraph did not think the subject. By producing a confident paragraph about it, the thing proved, by its own output, that it was working with the shadow rather than the source.

This is not a criticism of language models as tools. It is a reminder that the boundary the Kena drew 2,500 years ago still holds, and that a civilization that forgets the boundary will confuse the map for the territory.

A young person at a chatbot screen unable to articulate an inner feeling

Consider something more intimate. Think of a moment when you said 'I love you' to someone and felt, as you said it, that the sentence was pointing at something it could not contain. You were not lying. You were also not adequately describing anything. The words were a door you were asking the other person to walk through. Every deep human communication has this structure. The words carry you to a recognition, and the recognition is not itself a word. The Kena says ultimate reality has this same structure, scaled up infinitely.

Consider physics. Physicists who have spent a lifetime on quantum mechanics often say no one understands quantum mechanics; they can only calculate with it. The mathematics is precise. The physical picture behind the mathematics is something no one has put into ordinary words without contradiction. This is the Kena's distinction between the two kinds of knowing. Pratibodha-viditaṃ matam. It is known through the awakening.

And consider your own inner life right now, as you read this sentence. The words are moving through an awareness that is not itself one of the words. You are the reader, and the reader is not a sentence in this lesson. The Kena's claim is that if you turn attention to that reader, you will not find a new object to examine. You will find the thing inside which all examination is happening. That finding is not a proposition. It is a recognition. And it is available in the current moment, in the middle of an ordinary life.

This is what the Kena has and the Tractatus does not. The Tractatus ends with a wall and asks you to be silent in front of the wall. The Kena ends with a pointer that says the wall is the front of your own mind, and the thing you are looking for is what is doing the looking.

Back in the prison camp at Cassino, Wittgenstein closed his manuscript on the line that became his epitaph. He had reached the boundary. He set down his pen. The Kena rishi, twenty-five centuries earlier, had also reached the boundary, and then taken one more step.

Key figures

The Rishi of the Kena Upaniṣad

Late Vedic period, traditional composition within the Talavakāra Brāhmaṇa of the Sāma Veda (mid to late 1st millennium BCE)

Ludwig Wittgenstein

1889 to 1951 CE, Vienna, Cambridge, and Norway

Ādi Śaṅkara

8th century CE (traditionally 788 to 820), born in Kaladi, Kerala

Case studies

The Yakṣa, the Grass Blade, and the Daughter of the Himālaya: The Kena's Own Story

In the third khaṇḍa of the Kena Upaniṣad, the text interrupts its own philosophical precision to tell a story. The gods, the story says, had just won a great victory over the demons (asuras) in one of the perennial cosmic battles of the Vedic imagination. They were becoming proud. They were beginning to think that the victory had been their own doing, that their powers were their own possession, that they had won because they were strong. A mysterious yakṣa, a radiant being whose form none of them could identify, appeared in front of the gods. They did not know what it was. They sent Agni, the fire god, to investigate. Agni approached the yakṣa and announced who he was. 'I am Agni, Jātavedas, I who know all things that are born.' The yakṣa placed a single blade of grass on the ground between them and said, 'Burn this.' Agni, whose nature is to burn anything, gathered himself and tried with all his power to burn the blade of grass. Nothing happened. The grass did not burn. He returned, shaken, to the other gods. They sent Vāyu, the wind god. Vāyu approached the yakṣa and announced, 'I am Vāyu, Mātariśvan, I who can blow anything.' The yakṣa placed the same blade of grass between them. 'Blow this.' Vāyu tried with all his power. Nothing happened. The grass did not move. He too returned shaken. Finally Indra, king of the gods, went himself. As Indra approached, the yakṣa disappeared. In its place Indra saw a woman, radiant with light, wearing ornaments of gold: Umā Haimavatī, daughter of the Himālaya. She spoke to him directly. 'That was Brahman,' she said. 'It was in the victory of Brahman that you were rejoicing.' The gods' victory had been Brahman's victory working through them, and they had mistaken the channel for the source.

The yakṣa story is the Upaniṣad's own dramatization of the philosophical claim it had just made in the first two khaṇḍas. The philosophical claim was that Brahman is the source of every act of knowing, speaking, and acting, and cannot itself be caught as an object of any of those. The story shows what happens when this claim is ignored. Agni and Vāyu approach the yakṣa as if it were an ordinary object they could name and handle. They announce their powers as if those powers were their own. They are humiliated, not because they are weak, but because they have misunderstood the structure of their own strength. The same fire that burns everything cannot burn a blade of grass when the condition of its burning is withheld. Indra is the only one who gets the teaching, and only because he approaches differently. The yakṣa disappears for him, which is to say that the Upaniṣadic encounter cannot happen when the seeker still expects to find an object. Umā then arrives and names the yakṣa as Brahman, but notice the structure: she does not hand Indra a concept. She names the yakṣa after it has disappeared. The naming is a confirmation of a recognition that has already happened in Indra's own awareness. This is exactly the structure of pratibodha that Kena 2.4 will describe in philosophical language. The story does in narrative what the verses do in analysis.

The yakṣa story became, over the next three thousand years, one of the most quoted illustrations in Indian philosophical literature. Śaṅkara analyzes it in detail in his Kena Bhāṣya. Later Advaita teachers use it as the paradigmatic example of how pride in personal agency collapses under Upaniṣadic inquiry. It is also one of the earliest places in world literature where the feminine divine (Umā Haimavatī) appears as the teacher of the masculine divine (Indra), a reversal the text treats as completely natural. And its structural lesson, that the teaching cannot be transmitted as a proposition but only through an encounter that changes the student's recognition, became the template for every later guru-disciple pedagogy in the Vedānta tradition. The story is the Kena's answer to its own question: how do you know Brahman? You do not know it as a fact you acquire. You know it as a recognition that comes when the assumption of personal doership has been loosened. Agni and Vāyu were not ready. Indra was. The story explains, without saying it, why the Upaniṣadic teaching has historically required a teacher and a student in the same room.

Every moment you are sure that the action you just took was your own doing, the Kena is watching. The victory that feels personal was not personal. The thought that feels like yours was not quite yours. The Kena is not denying that you are the one doing things. It is asking you to notice the condition under which your doing is possible, and to see that that condition is not itself something you are doing. The next time you are tempted to take credit for a success, or to take full responsibility for a failure, run the Kena check: try to burn the blade of grass. Try, with all your power, to do the thing you think you are the author of. Notice what holds up and what does not. The residue, after the check is run, is the place the Upaniṣad is pointing at.

Gödel in Vienna, 1931: A Formal System Proving Its Own Limits

On 7 October 1930, at a conference in Königsberg, a 24-year-old Austrian mathematician named Kurt Gödel stood up and announced, almost as an afterthought, a result that would transform logic, mathematics, and the philosophy of language for the rest of the twentieth century. The result, published the following year as 'On Formally Undecidable Propositions of Principia Mathematica and Related Systems', had two parts. The first part, now known as Gödel's first incompleteness theorem, showed that any sufficiently rich formal system, one strong enough to express basic arithmetic, must contain true statements that cannot be proved within the system itself. The second part, the second incompleteness theorem, showed that such a system cannot even prove its own consistency from the inside. Gödel's proof was technical and exact. He constructed a mathematical sentence G that said, in the language of the system, 'this sentence cannot be proved in this system'. If G can be proved, then G is false, which means the system proves a false statement and is inconsistent. If G cannot be proved, then G is true, which means there is a true statement the system cannot reach. Either way, the system is incomplete. The proof was a kind of self-referential mirror held inside the system, and what it showed was that every sufficiently powerful formal language contains a point at which it cannot fully account for itself. The Vienna Circle, which had been building its philosophical program on the assumption that all meaningful questions could in principle be answered by formal analysis, was shaken. One of its members, the logician Hans Hahn, described Gödel's result as 'the end of the dream'. The mathematician John von Neumann, hearing the proof, reportedly abandoned his own work in foundations of mathematics and never returned to it.

Gödel's theorems are not a mystical statement. They are a mathematical proof. But their structural similarity to the Kena's central claim is striking enough that several twentieth-century philosophers, from Jean Ladrière to Douglas Hofstadter, have drawn the connection. The Kena had said, 2,500 years earlier, that thought cannot fully account for thought, because the one doing the thinking can never become the object of the thinking without generating a contradiction. The Kena's version of this claim was philosophical and pointed toward Brahman. Gödel's version was formal and pointed toward mathematics. But the shape of the claim is the same: any sufficiently rich system that tries to describe itself from the inside will hit a boundary where self-description fails. For Gödel, the boundary was the undecidable sentence G. For the Kena, the boundary was Kena 2.3: 'he who thinks It, knows It not'. Both proofs are in a sense performative. Gödel shows that mathematics has a horizon by producing a particular sentence that sits on the horizon. The Kena shows that thought has a horizon by producing a particular sentence whose very form demonstrates the horizon in the reader who tries to think it. There is a meaningful difference. Gödel's horizon is inside a formal system and is surrounded by other formal systems; you can climb to a stronger system and prove G from there, but the new system will have its own G. The Kena's horizon is the horizon of propositional thought in general, and there is no stronger propositional system that you can climb to in order to cross it. The only move available is the move the Kena recommends: pratibodha, immediate recognition, which is not a proposition. Gödel proved that formal language has an inside-out limit. The Kena had already proved that thought has such a limit, and had proposed a positive response to the limit rather than leaving the seeker stranded at it.

Gödel's theorems ended the formalist program that had dominated early twentieth-century foundations of mathematics. They also, indirectly, vindicated Wittgenstein's Tractatus and the broader tradition of recognizing that language has intrinsic limits. Within a decade of Gödel's paper, the entire landscape of mathematical philosophy had shifted. Within a century, the same shape of result has appeared in other domains. Alan Turing's halting problem (1936) is a Gödel-shaped result in computability. Chaitin's constant is a Gödel-shaped result in algorithmic information theory. The contemporary debate about whether large language models can represent their own reasoning to themselves is a Gödel-shaped question in artificial intelligence. In every case, the pattern is the same: a sufficiently rich self-referential system hits a limit at which it cannot fully describe itself. The Kena looked at this pattern with different tools and saw that the limit is not a failure of the system. It is a doorway to a different mode of knowing. Every modern reader who has been surprised or disturbed by Gödel's result is standing in a much older room than they realize.

The limits of thought are not a punishment. They are a diagnostic. When you hit a question you cannot think your way out of, the instinct is to try harder thinking. The Kena is saying, and Gödel's theorems confirm in a completely independent way, that some questions are not failures of effort. They are structural. A formal system cannot prove its own consistency from the inside. A mind cannot think its own thinker as an object of thought. The correct response is not more effort on the same level. It is a change of level. Gödel changed levels by constructing a meta-mathematical proof in ordinary language about what formal systems can do. The Kena changes levels by dropping the attempt to think Brahman and recognizing, instead, the awareness in which every thought is arising. The two moves are very different, but they share a single insight: when a system hits its own boundary, the way forward is not through the boundary but up, into a new mode of recognition that was always available and was never an object of the system in the first place.

Historical context

Late Vedic period for the Kena Upaniṣad (mid to late 1st millennium BCE); Early twentieth century Europe for Wittgenstein's Tractatus (1918 to 1921) and for Gödel's incompleteness theorems (1930 to 1931)

The Kena Upaniṣad was composed in the late Vedic period inside the Jaiminīya (Talavakāra) recension of the Sāma Veda. Unlike the Bṛhadāraṇyaka or the Chāndogya, the Kena is not part of an extensive narrative framework. It is short, concentrated, and almost austere, which is why it occupies the place it does in the Upaniṣadic curriculum: it is often taught after a student has already encountered the cosmological and ethical Upaniṣads and is ready for the purely epistemological question of how knowledge works at all. The Kena's treatment of language and thought presupposes a cultural context in which mantra (chanted speech), vāc (spoken word), and jñāna (knowledge) were already being debated as central categories of Vedic philosophy. The Upaniṣad's intervention was to draw a surgical line inside that debate: all of mantra and vāc and jñāna is real, is important, is useful, but none of it is Brahman.

Two civilizations, separated by 2,500 years, produced independently the same result about the limits of language and thought. The Kena produced it in the idiom of Vedic philosophy and linked it directly to a positive practice of recognition. Wittgenstein produced it in the idiom of formal logic and left it at the boundary without a practice. Gödel produced it in the idiom of pure mathematics and showed that the limit was not just philosophical but provably structural. That three such different traditions, using three such different tools, converged on the same insight is the strongest kind of evidence there is that the insight is real. The Kena has the additional credential of having arrived there first by a very large margin, and of having done more with the result than either of the twentieth-century thinkers was able to do.

Living traditions

The Kena Upaniṣad's claim that language cannot catch its own source is, in the twenty-first century, one of the most contemporary pieces of Sanskrit literature in existence. It shaped Śaṅkara's entire epistemology of Advaita. It shaped the modern self-inquiry method of Ramana Maharshi, who without having read the Upaniṣads taught exactly the move Kena 2.3 describes. It shaped the twentieth-century Indian philosopher J. Krishnamurti, whose lifelong insistence that 'the word is not the thing' is a direct echo of Kena 1.4. And it is increasingly cited in contemporary philosophy of mind, philosophy of language, and the philosophy of artificial intelligence, because the question the Kena asks is the question the twenty-first century is forced to confront every time a sufficiently fluent machine produces a sentence about a subject it does not understand. Every debate about what large language models are or are not doing is a debate whose terms the Kena already laid out. The Upaniṣad is 2,500 years old and is, in the strict technical sense, more current than it has ever been.

Reflection

More in What Can We Know? The Limits and Nature of Knowledge

All lessons in What Can We Know? The Limits and Nature of Knowledge · The Big Questions: Upanishads and the Philosophers Who Followed course