Can the self ever become one more thing among the things it knows?
न दृष्टेर्द्रष्टारं पश्येः, न श्रुतेः श्रोतारं शृणुयाः, न मतेर्मन्तारं मन्वीथाः, न विज्ञातेर्विज्ञातारं विजानीयाः । एष त आत्मा सर्वान्तरः ।
You cannot see the seer of seeing. You cannot hear the hearer of hearing. You cannot think the thinker of thinking. You cannot know the knower of knowing. This is your self, the one within all.
01 SHOW ME THE SELF
Uṣasta is dissatisfied.
He has asked Yājñavalkya to explain the brahman that is immediate and not indirect, the self that is within all. Yājñavalkya initially points toward the functions of life: that which breathes through breathing, that which acts through the different movements of breath.
Uṣasta refuses the answer. His objection is almost aggressively concrete. In editorial paraphrase: this is like telling me, there is the cow, there is the horse. You have pointed. You have named. But you have not shown me what I asked for.
The scene matters because Uṣasta wants something stronger than a definition. He asks for what the Sanskrit characterizes as sākṣād aparokṣād, something immediate, not known merely at a distance or through another thing.
The debate appears within a larger sequence of intellectual contests in the Bṛhadāraṇyaka associated with Yājñavalkya. The approximate c. 7th–6th century BCE placement concerns broad early textual material; it does not date this narrated conversation as a historical event. Uṣasta’s demand appears reasonable: if you claim that there is a self, show it.
Yājñavalkya’s response is remarkable because he does not finally point to anything. He questions the demand itself.
02 THE SEER OF SEEING
Yājñavalkya begins with vision. Not: you cannot see the self. Something more unusual: you cannot see the seer of seeing.
The construction matters. There is sight. There are things seen. And then there is whatever it is by virtue of which seeing occurs.
Normally our attention moves outward. I see a tree. A face. A page. A hand. Even when I turn inward, I may notice sensations, memories, emotions or thoughts. Those too can become things that I observe.
But Yājñavalkya introduces another question: what about the one to whom they appear? If I find another object, have I found the seer? Or merely something else seen?
The same structure is then repeated. You cannot hear the hearer of hearing. You cannot think the thinker of thinking. You cannot know the knower of knowing.
The repetition is not decorative. Each line closes another possible escape. Perhaps the self is something visible. No. Perhaps audible. No. Perhaps it is a thought. Again no. Perhaps it can simply be known like any other fact. The final line blocks that too.
The passage constructs a distinction between what can become an object of awareness and that to which objects become present. Modern philosophical language may call this a distinction between object and subject. This is a later interpretive vocabulary. The Sanskrit does not itself use modern subject/object terminology.
03 THE PROBLEM WITH LOOKING
Suppose I decide to find my self.
I begin with the body. I can see my hand. I can feel my breathing. I can observe pain. These therefore appear to me.
Perhaps the self is memory. But a memory can appear and disappear. I can remember yesterday and examine what I remember.
Perhaps the self is a thought. But I can notice the thought. I can disagree with it. I can wait for it to pass.
Perhaps the self is a feeling of being me. But even that feeling can itself become something noticed.
The structure produces a strange regress. Every time I say: there it is. Yājñavalkya can ask, in our reconstruction: and what knows that?
A modern reader might describe Uṣasta’s demand as risking a category mistake: trying to encounter the subject as though it were simply another object. This is a modern philosophical reconstruction, not Yājñavalkya’s terminology.
Yājñavalkya’s own method is more concrete: seeing, hearing, thinking, knowing. The philosophical movement emerges from verbs.
04 ĀTMAN IS NOT A PERSONALITY PROFILE
The sentence eventually names what it has been circling: ātman. English usually gives us: self. That translation is necessary and dangerous.
Modern readers hear “self” and may imagine personality, preferences, memories, history, identity, the psychological story someone tells about who they are. But none of those is what this passage has isolated. They are all things that can, at least in principle, become objects of reflection.
The Bṛhadāraṇyaka participates in an older Upaniṣadic inquiry into ātman and its relation to knowledge, reality and brahman. This inquiry should not be collapsed into one later systematic Vedānta doctrine, or made to stand for Hinduism generally.
Later reception. Śaṅkara’s Advaita belongs to a later interpretive history. We should not silently read it backwards into the early text. A later systematic account may help us ask questions, but it cannot simply be substituted for this passage.
05 CAN THE KNOWER BE KNOWN?
The final formulation is the most difficult: you cannot know the knower of knowing.
It sounds self-defeating. If the knower cannot be known, how can Yājñavalkya tell us anything about it? How can he even say that it exists?
This is where the passage becomes philosophically interesting. There is more than one possible meaning of knowing. One form of knowledge places something before us. We identify it. Describe it. Compare it. Classify it. In that form, knowledge takes an object.
Yājñavalkya may be denying that the self can be known in that way. The denial does not necessarily imply complete ignorance. It may imply that the relation we have to the self cannot simply reproduce the ordinary relation between knower and known.
This returns us to Uṣasta’s demand. He wants direct knowledge. He expects directness to mean: show it to me. Yājñavalkya suggests another possibility: perhaps nothing is more immediate than that by which showing is possible.
The self may be difficult to find not because it is infinitely distant, but because it is too close to stand before us.
06 THE ONE WITHIN ALL
The selected passage concludes: eṣa ta ātmā sarvāntaraḥ. This is your self, the one within all.
“Within” need not be understood spatially. The phrase complicates any purely psychological reading. Yājñavalkya is not simply telling Uṣasta to search deeper inside a private personality.
One interpretive possibility is that “within” describes a structural relation rather than a physical location. Seeing cannot occur without a seer. Hearing without a hearer. Thinking without a thinker. Knowing without a knower.
This structural-inwardness interpretation is a Paragraphos Reading of sarvāntaraḥ, not a claim that its lexical meaning settles the philosophical question. Nor does the sentence by itself license replacing ātman with “universal consciousness”. That would require further argument and historical qualification.
07 THE THING YOU CANNOT PUT IN FRONT OF YOU
There is a familiar habit in philosophy. Turn a problem into an object. Define it. Separate it. Inspect its properties. The method is powerful.
Yājñavalkya’s answer introduces a limit case. What happens when the thing under investigation is precisely that from which investigation occurs?
This argument does not by itself prove an eternal soul, Advaita, identity with brahman or any complete later metaphysical system. It does not settle later Indian debates about selfhood.
Comparisons with Buddhist critiques of self need particular texts and arguments. The simplification “Hinduism believes in self, Buddhism does not” is unacceptable: it substitutes two uniform positions for distinct inquiries and disagreements. Our connection is an editorial comparison, not a claim that early Buddhism directly answered Bṛhadāraṇyaka 3.4.2.
Return to the text.
See something.
Then try to see the seer.
Hear something.
Then try to hear the hearer.
Notice a thought.
Then try to place the thinker beside the thought.
Know something.
Then try to make the knower one more known thing.
The failure may be the beginning of the teaching.
At first, we assume the self is difficult because we have not yet looked deeply enough. By the end, another possibility appears.
Perhaps we fail to find the self among objects not because our search has been insufficient, but because we have misunderstood what kind of thing we are looking for.
Sources / Apparatus
Primary text
Yājñavalkya, as represented in the Upaniṣad, Bṛhadāraṇyaka Upaniṣad. 3.4.2.
The supplied excerpt is retained against V. P. Limaye and R. D. Vadekar, Eighteen Principal Upaniṣads, vol. I (Poona, 1958), p. 218, Kāṇva 3.4.2, visually checked. Otto Böhtlingk's independent Mādhyaṃdina edition (St. Petersburg, 1889), p. 27, corresponding 3.5.1, supports the disputed verbal forms. This is a comparison of identified printed editions, not a fresh manuscript collation. Punctuation, unaccented spelling and five reading lines remain editorial.
Original Paragraphos working translation supplied for this publication. It is not attributed to a published translator.
This excerpt ends at eṣa ta ātmā sarvāntaraḥ; it omits ato 'nyad ārtam and Uṣasta's subsequent silence, so it does not present all of 3.4.2. Limaye–Vadekar and Böhtlingk print शृणुयाः and मन्वीथाः. Madhavananda's 1950 edition, p. 469, prints शृणुयात् in the second clause: a real printed discrepancy, probably a printing error, though its cause is not established. No manuscript support or intentional normalization for that reading was identified in the consulted material. The display follows the attested second-person शृणुयाः; शृणुयात् would be third person, not a sandhi equivalent. GRETIL has connected manvīthā, while the displayed मन्वीथाः has direct print support and follows a paused-clause convention. Grammar explains that distinction without proving GRETIL's editorial intention. GRETIL separates the root text from a later commentary attributed to Śaṅkara and warns that the commentary is not proofread. The broad early-material date is not a date for the narrated conversation.
The commissioning editor supplied and authorized publication of this Sanskrit excerpt and original working translation. No modern edition's apparatus or third-party translation is reproduced. This authorization covers the supplied material only.
Translation / Textual reference
- V. P. Limaye and R. D. Vadekar, eds., Eighteen Principal Upaniṣads, vol. I, Gandhi Memorial Edition (Poona: Vaidika Saṃśodhana Maṇḍala, 1958), p. 218, Kāṇva 3.4.2.
Title, publication statement, editorial method and Sanskrit passage visually inspected. The four negative clauses, both disputed verbal forms and selected ending support the retained display. The following phrase and Uṣasta's silence are outside the excerpt. This scholarly edition has selective comparative notes; neither its apparatus nor this comparison establishes manuscript unanimity. GRETIL names this edition among its checking sources, so the digitization is not counted as another independent witness.
- Otto Böhtlingk, Bṛhadâraṇjakopanishad in der Mâdhjaṃdina-Recension (St. Petersburg, 1889), pp. 26–27, Mādhyaṃdina 3.5.1.
Title, preface, passage and relevant critical discussion visually inspected. Page 27, lines 2–3, prints शृणुयाः and मन्वीथाः. Its preface identifies Weber's printed text and Berlin manuscript 1470 among its bases. This is an independent earlier printed comparison, not a manuscript inspected by Paragraphos. The Mādhyaṃdina section order and verse division differ from Kāṇva 3.4.2; the corresponding Uṣasta passage is 3.5.1 here.
- Swami Madhavananda, The Bṛhadāraṇyaka Upaniṣad with the Commentary of Śaṅkarācārya, third edition (Advaita Ashrama, 1950), p. 469, 3.4.2.
Visual inspection confirms मन्वीथाः and the differing शृणुयात्. The printed corrigenda, p. xxii, do not correct this word. A probable printing error is an inference from the other editions and the parallel second-person constructions, not an attested erratum or established manuscript variant. No third-party English or later commentary is reproduced or imported into the early argument.
- William Dwight Whitney, Sanskrit Grammar, §§170 and 177: final s and -ās.
The rules distinguish final visarga at a pause from loss after long ā before a voiced sound. This accounts for a possible paused/connected distinction between manvīthāḥ and manvīthā; it does not establish an edition's intention or textual history. The retained forms are separately attested in the inspected prints.
- GRETIL, Bṛhadāraṇyaka Upaniṣad, Kāṇva recension, chapter 3: root text 3.4.1–2.
Consulted for Uṣasta's demand, the breath functions, cow/horse comparison and four repeated prohibitions. The displayed passage is an excerpt. The electronic text reports comparison with Limaye and Vadekar's edition, now directly inspected. GRETIL alone is not the basis for the Sanskrit display.
- Working glosses and interpretive limits: sākṣād aparokṣād, draṣṭṛ, ātman, sarvāntaraḥ.
The terms occur in the located text; the selected Sanskrit has now been compared with identified printed editions. Further lexical scholarship may refine these working glosses. ‘Within’ as a structural relation and modern subject/object language are the present article's interpretations, not literal lexical definitions. Print agreement does not certify those interpretations.
Secondary reading
- Brian Black, ‘The Upaniṣads’, Internet Encyclopedia of Philosophy.
Consulted for the diversity of early Upaniṣadic inquiry, Yājñavalkya's debates and self-knowledge. Its broad chronology of Upaniṣadic composition does not independently date 3.4.2 to the seventh or sixth century BCE.
- Dating of the early Bṛhadāraṇyaka material.
The displayed c. 7th–6th century BCE is an approximate editorial period. Passage-specific dating and the history of this textual layer require specialist review; no exact year or date of an actual debate is asserted.
Reception
- ‘Śaṅkara’, Stanford Encyclopedia of Philosophy.
Consulted to distinguish Śaṅkara's later Advaita framework from the early source. His activity belongs roughly to the eighth century CE. No claim is made here to have established his passage-level reading of 3.4.2.
- ‘Buddha’, Stanford Encyclopedia of Philosophy.
Background for the comparison with arguments about non-self, not evidence of a direct response to this passage. Folio 0005 now reads SN 22.59; no historical influence claim is made.