The advice everyone gives is: find someone further along and stay near them.
This tradition agrees that proximity is the mechanism — association is the first mover of the whole path — and then adds two conditions that most people never think to check.
The person has to be further along. They also have to be fond of you. And, hardest to arrange, they have to be the same kind of devotee you are.1
Miss any one and the transmission doesn't happen. And the tradition is specific about what goes wrong in each case.
Sajātīya — of the same kind, the same species.
The requirement is that you associate with devotees who worship the same form of the Lord as you do, and who possess the same internal mood. Spelled out: those in the servant's mood should associate with devotees in the servant's mood; those in the friend's mood with friends; and similarly for the parental and amorous moods. Worshippers of Kṛṣṇa with worshippers of Kṛṣṇa; worshippers of a particular incarnation with those who worship the same form.2
This is a strange requirement if you think transmission is about level. It's obvious if you think it's about contagion of a specific thing. You don't catch generic advancement. You catch the particular mood the other person has, because that is what they have.
Which means the great devotee in a different mood is not a shortcut. Their company will not produce your mood, because their mood is not yours — and the book's whole account of the five relationships is that they are genuinely different relationships, not levels of one.
Snigdha — literally oily, smooth, and by extension soft-hearted, affectionate.
The gloss is precise about what it excludes: those who are harsh, unsympathetic, or indifferent. And the reason given is not about comfort:
The esoteric mysteries of bhajana and genuine instruction regarding the method of performing bhajana cannot be obtained from sādhus who are indifferently disposed or who are not affectionate.3
So affection is not a pleasant extra. It is the channel condition. The confidential material does not travel across indifference — not because the indifferent person is withholding, but because that kind of content apparently cannot be transmitted on a cold line.
Which is a substantial claim about pedagogy and it is stated once, in passing, as a criterion for choosing company. A tradition that spends sixty-four limbs on method here says that the real method is only available inside a relationship where somebody is fond of you.
Mahattara — greater, more.
And the specification names three axes: more advanced in steadiness in practice, in direct experience and realisation, and in knowing the confidential mysteries of scripture — in other words, superior to oneself in all respects.4
That's the least surprising of the three tests, and it's worth noting what it isn't. It isn't seniority, position, or reputation. Two of the three axes are internal and one is scholarly. None of them is institutional.
The book gives the failure mode plainly, and it's the crystal.
Quoting a verse: just as a crystal reflects the colour of objects that are brought near it, a person's nature will be exactly in accordance with the association he keeps. Therefore, by associating with pure devotees one becomes pure.5
And then the consequence for skipping the third test: by failing to associate with those more advanced, the advancement of bhakti is checked and one's disposition or nature will be on exactly the same level as those devotees whose association one keeps.
Not damaged. Levelled. You come to rest at the altitude of the room.
That's a more unsettling model than the usual one, because it has no floor and no ceiling of its own. The crystal has no colour. Whatever is nearest is what you are.
There's a small hermeneutic move here that I think is the most revealing thing in the passage.
The scriptures contain an instruction to live without association — niḥsaṅga. It's a standard renunciant ideal: no entanglements, no company, alone.
The book's reading: the actual purport of the instruction given in the scriptures to live devoid of association is that one should live in the company of sādhus.6
So be without company is read as be in the right company. The negation is converted into a specification.
You can call that a stretch, and it is one. It's also forced by the architecture: if the whole causal chain runs through association, then a scriptural command to have none of it has to be reinterpreted or the system breaks. And the tradition performs the reinterpretation in a single sentence, without argument, as though nothing were happening.
What a tradition does with the verses that contradict its mechanism tells you where its mechanism actually is.
You've found someone impressive. Deeper than you, clearly, and generous with time.
Run the three tests. Further along — yes. Fond of you — probably. Same disposition — and here you stop, because you've never thought about it, and when you do you notice their whole orientation is contemplative stillness and yours is something closer to wanting to be useful to somebody.
Under this scheme that mismatch is not a minor friction. It's a failed test, and what you'll absorb from them is their mood rather than the deepening of yours.
Now the harder one. The person who meets all three is available and you find them slightly embarrassing — less articulate than the impressive one, socially unimpressive, warm in a way that makes you uncomfortable.
The tests do not include impressive. They include affectionate, which the impressive person may not be, and same-disposition, which is invisible on first meeting and is the one people never check.
And the third application, which is a diagnosis rather than an action. If you have been in the same room for three years and nothing has moved, the crystal model says you have levelled at the altitude of that room. Not failed. Matched. Which is a different problem and has a different fix.
The three tests are usually quoted as a list, and reading them as a conjunction changes what they say.
Take each pair and the missing third names a recognisable failure.
Advanced and affectionate, wrong disposition. The warm, deep teacher whose mood isn't yours. You will get real transmission of something, and it will be their thing, and years later you will find you have carefully cultivated a relationship with God that isn't the one you actually have.
Advanced and same disposition, not affectionate. The right person, cold. On the book's own account the confidential material simply doesn't cross. You'll get the architecture and not the interior — which is, incidentally, an exact description of what you can get from a book.
Affectionate and same disposition, not advanced. A friend. Genuinely good for you, and the crystal levels you at their altitude, which is your own.
Three tests, three distinct failures, and the tradition's insistence on the conjunction is what makes finding company hard. Which may be the point: the difficulty is a feature of the specification, not a shortage in the world. A criterion this narrow explains why a tradition built on association also has an enormous literature about how rarely the right association is found.
Strongest support. Each test comes with its own named failure mode, and the affection requirement is given a mechanism rather than a sentiment — the confidential material does not travel on a cold line. The crystal verse is quoted from an external source rather than asserted.
The tension the book does not close. The sajātīya test requires you to know your own mood, and the book states elsewhere that a practitioner cannot conceive of their perfected identity without the mercy of a teacher — and that the identity arises only by the merciful indication of that teacher. So you must match on a property you cannot establish without first having found the person you were supposed to match with.
A second tension. Niḥsaṅga is read as its opposite in one unargued sentence. Whether or not the reading is defensible, it is the clearest instance in this book of a verse being turned to fit the mechanism, and it is done without flagging that anything has been done.
Open question. Does the crystal model have any asymmetry? As stated, it is neutral: you take the colour of what is near. But the tradition also holds that devotion is transmitted while ordinary conditioning is not — an advanced devotee sitting among the worldly doesn't level downward. Nothing here explains why the crystal works in only one direction.
The vault's Bryant material on association makes the case that spiritual company is a structural cause rather than a moral nicety, and reading it you come away convinced of the mechanism in general.
Come here and the mechanism is specified to the point of scarcity. Not company: this company, matched on mood, warm toward you, ahead of you on three internal axes. The general claim becomes an intake requirement, and most available company fails it.
That shift matters because it changes what the doctrine is doing. Stated generally, association is encouragement — go find good people. Stated with three conjoined tests, it is a filter, and the tradition's own literature about the rarity of genuine association stops looking like humility and starts looking like arithmetic.
Match the source to your inclination, not to their eminence — because what transmits is their specific thing, not their altitude. Set this against Choose the Mentor by Inclination. That page's argument is that the right teacher is selected by fit with your own bent rather than by their standing, and that the eminent mismatched mentor actively costs you — you spend years acquiring somebody else's aptitude.
Sajātīya is that principle with the mechanism made explicit: you absorb the mood the other person has, because a mood is what they've got to give, and the five moods are genuinely different relationships rather than rungs.
Here's what the pair produces that neither gives alone. Both traditions independently discovered that the matching test is the one nobody applies, and both explain why: eminence is legible on first contact and inclination isn't. You can see that somebody is further along in about ten minutes. You cannot see whether their bent is yours, and — this is the part the devotional version adds — you may not yet know your own. Which makes the matching test structurally the last one you can run and the first one that matters, and that ordering is why the failure is so common and so slow to surface. It shows up as years of well-conducted work in a register that was never yours, and it looks like progress the whole way, because it is progress. Just not yours.
Sharpest implication. Three tests, held simultaneously, and the one people never check is the match. Eminence is visible immediately, warmth is visible soon, and disposition is invisible — possibly to both parties — which means the criterion that decides whether the years will count is the one you are least able to evaluate at the moment you commit.
Generative questions.