You asked somebody for a thing and they did not give it to you, and later you found out they had it the whole time.
That is one situation. Here is the other one, which feels identical from inside and is not the same at all: you asked somebody for a thing, they had it, they did not give it to you, and they were right. The coach who will not tell you the trick that would have got you through the routine this month, because you would then have a routine and no technique. The friend who will not lend you the money, having watched what the last loan did. The manager who declines to solve it for you.
In every one of those cases the refusal is doing something the granting could not do, and in every one of them the person refused experiences it as being let down by someone who could have helped.
You cannot tell the two situations apart at the time. Not from where you are standing, not with the information you have. The only thing that separates being helped and being stiffed is the other party's reasoning, and you do not have access to it.
Madhya 22 has God make that argument about himself, out loud, in the first person, about prayer.
The claim arrives as a piece of quoted speech, and it is startling how blunt it is.
"Since I am very intelligent, why should I give this fool material prosperity? Instead I shall induce him to take the nectar of the shelter of My lotus feet."1
Note what that sentence is not. It is not I withhold prosperity because prosperity is bad for you — that argument is everywhere in this literature and it is not the one being made. It is not compassion, it is not discipline, and it is not a test.
It is stated as a competence claim. The reason given for the refusal is that he is intelligent. The prayer is described as the request of a fool, and the refusal as what a clever party does when a foolish party asks for something.
Which is a genuinely unusual move for a devotional text, and it is worth pausing on the shape of it before the content. Most traditions defend unanswered prayer on grounds of love, mystery, or a larger plan. This one defends it on grounds of skill. The petitioner has made a bad request; the recipient is good at this; therefore the request is overruled.
And there is a second line, which supplies the actual rule.
"He does not award anything which, after being enjoyed, will cause someone to petition Him again and again."2
That is not an attitude. That is a filter with a stated criterion, and the criterion is not whether the thing is good, or deserved, or harmful. It is whether granting it produces another request.
Sit with that filter, because it is the page.
Two people ask for the same thing. One of them will be satisfied by it and stop; the other will get it, enjoy it, and be back next year for the larger version. On the stated rule, the first gets it and the second does not — and the thing requested was identical.
So the doctrine is not about objects at all. It is about the shape of the demand curve behind the request, and it sorts petitions by what happens after they are granted rather than by what is being asked for.
Which is a real and recognisable thing. Some wants terminate on satisfaction. You want a coat, you get a coat, the wanting stops. Others are structurally unsatisfiable — the want reconstitutes at a higher level on being fed, and the object was never the point. Anybody who has watched somebody chase a number knows the second kind, and knows that granting it is not kindness.
The corpus is claiming that God can tell which is which, in advance, per person, and routes accordingly.
And notice what that does to the petitioner's evidence. Under this rule, being refused tells you something about yourself: you asked for a thing whose granting would have kept you asking. The refusal is a diagnosis delivered as a non-event. You are not told; you simply do not receive, and the doctrine says the not-receiving was the message.
There is no way to run this backwards. A person who prayed and got nothing cannot distinguish my request was of the recurring kind from nobody was listening, and the corpus supplies no third reading.
Ādi 8 supplies the other half, and it is stranger.
Ask for sense gratification, the corpus says, and Kṛṣṇa delivers it immediately. Ask for liberation, and he delivers that too. But pure devotional service he keeps hidden — and the reason given is not that you are unready.
The reason given is that by devotional service he is bound to the devotee.3
Read that again with the roles in view. The thing being withheld is withheld because giving it would put the recipient in a position of power over the giver. This is not the intelligence argument from Madhya 22, where the refusal is for the petitioner's benefit. This is a party protecting his own freedom of action, stated plainly, in a text whose whole purpose is to recommend the withheld item.
And the two doctrines sit unhappily together. Madhya 22's God refuses what would keep you dependent. Ādi 8's God refuses what would make him dependent. Both are called mercy in their surrounding purports. Only one of them has the petitioner's interest anywhere in it.
The corpus does not put these on the same page and does not notice the difference. Both are supplied to the same reader, who is being urged, in the same chapters, to ask for exactly the thing that is being withheld on the grounds that giving it costs the giver something.
Now line up what gets granted against what the corpus says is worth having, because the two lists are in opposite order.
Material enjoyment — the lowest item in every valuation this tradition makes, filed among the weeds, described as the fool's request. Granted immediately on asking.
Liberation — the terminal goal of most of Indian religion, and this corpus's second-tier attainment, disparaged but real. Granted immediately on asking.
Pure devotional service — the summit, the thing the sixty-two chapters exist to recommend, the only item the tradition calls the actual point. Withheld.
So the availability schedule is an exact inversion of the value schedule. The cheaper the item, the faster the delivery. The one thing you are told to want is the one thing the doctrine says you will not simply be given for asking.
Which is coherent — a thing that cannot be had on demand is by that fact not a commodity, and the tradition wants devotion to be uncommodifiable. But it has a consequence the corpus does not draw. If asking works for everything except the thing worth having, then prayer is a mechanism that functions in inverse proportion to the importance of the request, and a practitioner following the instruction to ask is being pointed at the one channel the doctrine has declared closed.
The corpus's answer, implicitly, is that devotion arrives by grace rather than by request. Fine. But it never says what the petitioning is for, in a system where the good stuff does not come through petitioning.
Here is the structural problem, and this build has now met it three times in three different chapters.
Under this doctrine, there is no possible outcome that counts as evidence against it.
You ask and receive: he grants requests. You ask and do not receive: he is too intelligent to grant that one, and the refusal is the higher gift. You ask for the highest thing and do not receive: it is being kept hidden, deliberately, for reasons internal to him. Every branch confirms.
That is the same shape as two other tests in this corpus. Absence of feeling after long chanting proves the practitioner's offence rather than the method's limit. Convincing symptoms in the wrong person are declared counterfeit by an expert.4 Efficacy, authenticity and now responsiveness are each tested by an instrument that cannot return a negative.
The honest thing to say is that this is not a lapse. It is what a doctrine looks like when it is built to survive the ordinary experience of its practitioners — which is that most prayers are not answered — and a tradition that could not absorb that fact would not have lasted five hundred years.
But it costs something specific, and the cost is worth naming. A promise that cannot be broken is also a promise that cannot be relied on, because reliance means knowing what would count as the promise failing. Under this doctrine the devotee has an assurance and no expectation, and those are different possessions.
Layer A gives the two lines compactly — the intelligence argument in Madhya 22, the hiding of bhakti in Ādi 8 — inside instruction, without commentary.
Layer B expands both considerably. The purports around the Madhya 22 material press the point into a general theory of how the Lord manages his devotees' material circumstances, with contemporary examples.5
Worth marking the direction of the expansion. The verse says he does not award what produces repeat petitions. The commentary tends to say he actively removes what a devotee already has. Those are different claims — one is a rule about granting, the other is a doctrine of dispossession — and the second is considerably harsher than what the verse supplies. The vault holds that harsher version separately, from another source, and it is worth reading against this one rather than merged into it.
Nothing in the range read has Kṛṣṇadāsa make the dispossession claim in his own voice.
You asked for something reasonable and did not get it, and you are trying to work out what that means. Notice that you have no instrument for this. The refusal-was-wisdom reading and the nobody-was-listening reading produce identical evidence, and choosing between them is a decision about what to believe, not a finding. Make it as a decision.
You are the one refusing. Somebody has asked you for the shortcut, the loan, the answer. The corpus's criterion is a good one and it is not about the thing asked for — it is about whether granting it ends the asking or restarts it. Run that test, and then say which one you ran, because the person in front of you cannot see it and will otherwise read the refusal as indifference.
You are being told that not getting the thing is itself the gift. Ask what outcome would have counted as not-a-gift. If there isn't one, you are being handed reassurance rather than information, and it is worth knowing which of the two you have.
Somebody withholds a thing from you and gives a reason about your own good. Check whether granting it would have transferred something to you at their expense. Both motives can be true at once, and the second is the one nobody states.
The evidence is two short passages, thirty-one chapters apart, each stated as a general rule rather than as a story.6 Neither is argued for; both are asserted in Kṛṣṇa's own voice, which is the corpus's strongest register.
The tension the source doesn't close: two incompatible reasons for the same refusal. Madhya 22 withholds for the petitioner's sake; Ādi 8 withholds because granting would bind the giver. The corpus calls both mercy and never sets them beside each other.
A second one. The corpus urges the reader to want pure devotion and states that pure devotion is the one thing not distributed on request. The instruction and the mechanism point in different directions.
Open: what petitioning is for in this system. If the low items are granted and the high one is withheld, the corpus never says what a devotee should be asking for, and nothing in the range read addresses it.
Against The Plundered Storehouse this is the corpus at its most guarded. There, love is broken open and distributed with no test of fitness and no conditions whatever; here, the same commodity is described as deliberately concealed for reasons internal to the giver. The most generous page in this corpus and the most withholding one are about the same substance.
It supplies the missing motive for Refusing the Five Liberations. That page has devotees decline liberation when offered; this one records that liberation is handed over on request while devotion is not. Liberation is cheap in this economy — both freely given and freely refused — and the two pages are the supply side and the demand side of one valuation.
And it stands beside Kṛṣṇa as Debtor as a straight contradiction worth keeping open. That page has him unable to clear an obligation to the gopīs, permanently in their debt by his own admission. This one has him withholding devotional service precisely so as not to be bound. He avoids the bond as policy and confesses to it as fact, and the corpus reports both as glories.
Intermittent Reinforcement and Control Dynamics — that page's mechanism is the most reliable engagement engine known: reward a behaviour some of the time, unpredictably, and you get sustained, intense, extinction-resistant repetition. The slot machine, the inconsistent parent, the partner who is sometimes warm. Crucially, the mechanism does not care about the source of the unpredictability — accident, cultivation and engineering produce identical effects.
Put the two side by side and something specific emerges, which neither produces alone.
The corpus's stated rule is the exact inverse of intermittent reinforcement. Kṛṣṇa, it says, "does not award anything which, after being enjoyed, will cause someone to petition Him again and again." That is a declared policy of not generating repeat requests — an anti-slot-machine, deliberately designed to terminate the petitioning loop rather than sustain it.
And the behaviour it produces is the behaviour intermittent reinforcement produces. A practitioner who prays, sometimes receives, often does not, cannot predict which, and continues indefinitely. Requests granted immediately at the low end; the high request never granted at all; no rule available to the petitioner for telling in advance which is which.
Which is the insight: a stated motive and a produced effect are independent, and a doctrine can disclaim the mechanism it is running. The reinforcement literature reads unpredictable reward as a control device and looks for the controller's interest. This corpus supplies a variable-outcome system with an elaborate account of why the variability is benevolent — and the account is unfalsifiable, so it cannot be checked against the alternative.
The practical residue is not a verdict on this tradition. It is a test: when a system delivers unpredictably and explains the unpredictability, ask whether the explanation would look different if the unpredictability were serving the deliverer. In a great many cases it would not, and that is precisely when the explanation stops carrying information.
Sharpest implication. A doctrine of the deliberately unanswered prayer buys enormous resilience — no outcome can embarrass it — and pays for that resilience in a currency nobody notices: the petitioner ends up with an assurance that cannot fail and an expectation that cannot form. The two feel like the same thing right up until you need to plan on one.
Generative questions.