Chapter 8 Eternity and creation¶

The theorem leaves its domain. The first seven chapters completed their arguments within a Buddhist setting. This chapter takes the theorem outside that setting for a test. The thesis is a combination of concepts: eternal, self-existent and not constituted by another, immutable, omniscient, and creating the world by free will. This chapter names no existing religion and cites no scripture. It addresses only the philosophical structure of this combination itself. Any form of faith that does not adopt this combination lies outside its scope. Every conclusion in this chapter is to be read conditionally. This is the application beyond Buddhism of Chapter 2's conditional clause: whoever accepts this schema bears this result. Whether a historical configuration accepts it is for that configuration's own account to determine; this chapter does not judge on its behalf. The reason for choosing this test case is structural. The combination writes D2's three conditions explicitly into its definition, then assigns this bearer of intrinsic nature the heaviest relational duties. If the theorem works, the contradictions must be densest here. After the five routes, we revisit three doors and then bring the account together.

Route one: a textbook case and a test of grammar. Self-existence, not being constituted by another, is an explicit definition of independence at the level of existence. Eternal immutability is the second condition. Being uncreated is the third. When the preceding chapters examined dharma entities, atoms, and goers, the conditions had to be extracted one by one from the opponent's usage. Here most are written in the definition's first line. But one point must be recorded precisely. The first line explicitly states independence at the level of existence; D2's first condition in its full scope is independence at the level of identity. How do the two levels connect? There are two routes. The mature forms of this combination have always come with a simplicity clause: no internal distinctions or composition, essence and existence being one. Once this clause is signed, independence reaches the level of identity and is accepted in full. If a configuration does not sign, accepting only existential independence while allowing its essence to depend contrastively on something else, it has already surrendered full independence under Chapter 2's conditional clause. It lies outside the scope of application, at the cost of failing to live up to its name of self-existence. E3 needs no dispute. Advocates of this combination would never accept its object as a mere designation. Applying the theorem therefore needs only one check: does this configuration accept full independence? If so, it enters the scope of application, and the mainstream forms have always accepted it.

The step marked in advance in Chapter 4 is now made. Let this self-existent being describe itself. To say that it is such and such, any predicate marks out a class. Marking out a class is contrast; contrast introduces what is excluded. Yet its self-existence expressly forbids its identity from depending on any difference. Grammar thus leaves it only one form of sentence: "I am I." The direction of explanation must be explicit. This chapter cites no traditional wording as evidence. It only states a theorem of grammar in advance: on an ontological reading, a self-existent being's account of itself must, and can only, take the form of a tautology. Content violates self-existence; keeping self-existence leaves only tautology. Which historical sentence fits this form, and what its correct reading is, belong to philology. Any self-description of this form admits two readings. The relational reading, "I will be with those addressed," lives entirely at the level of E2, mutually dependent, responsive to conditions, and functional. The theorem does not touch a word of it. The ontological reading, "the self-subsisting being," "existence itself," falls directly into the dilemma. The theorem thus supplies only a format for sorting the two readings. Which is the correct reading of which sentence is left to philology and reception history. This chapter has said nothing on that question.

Route two: permanence and action. Creation is an action. Action has a before and after: not yet acting, acting, having acted. What is eternally immutable has no before or after. It must be wholly the same before creation and at the moment of creation, or the second condition fails. Put the two together. If there is no difference on the agent's side between the moment of creation and before it, creation cannot be an action of the agent. At most it is an appearance of the world without any participation by the agent, and "creation" loses its content. If there is a difference on the agent's side, the permanent being has changed. The same inference has already run once in Essay 15's treatment of the Sarvāstivāda distinction between a dharma's entity and its function. The entity is permanent, while function arises and ceases. If function's arising and ceasing does not affect the entity, function has nothing to do with the entity. If it does affect it, the entity is not permanent. The argument repeats outside Buddhism without a word changed. Will cannot escape the same question. If an eternal will eternally wills creation, creation should be as eternal as the will. If creation has a beginning, at that beginning the will has turned once from not-yet to already. Such a turn is change.

This route must be assessed separately by configuration. For a temporal configuration, the preceding argument applies directly. An atemporal configuration adds a piece of doctrinal engineering: the relation of creation is real only on the created side, with no real relation on the agent's side, and hence no turn to speak of. This is the opponent's strongest form, and this chapter accepts it as stated. It evades the contradiction, and the cost is recorded at once. If there is no real difference whatever on the agent's side, the name "creator" is supported entirely by changes on the created side. Action and will no longer have a foothold on the agent's side. Under this configuration, route two yields no contradiction, only the price of idle spinning. The substantive inventory passes to route five.

Route three: one cause and all dharmas. This route has a precedent within Buddhism, the passage refuting Īśvara in fascicle 7 of the Abhidharmakośabhāṣya, cited in Essay 15, Chapter 5. If all dharmas arise from one cause, they should arise simultaneously. Yet we directly see them arising successively, so further conditions must be awaited. If further conditions are needed, one cause is insufficient as a cause. "What use is this Īśvara?" Thus "dharmas arising from a single cause are definitely nonexistent."¹ First make one distinction clear. Creating the world by free will does not mean being the sole cause of all dharmas. Created secondary causes can perfectly well account for succession within the world. The ancient refutation of a sole cause cannot be applied here unchanged. This route's target is the first moment. At the start of the whole sequence, no condition on the world's side is available for use. Only the eternal being itself is available. For configurations that also affirm omnibenevolence and require a reason for the world's beginning at this time and in this form, two alternatives arise: either there is no reason, contrary to this further requirement, or the reason lies within the eternal being, in which case the Kośa's objection concerning differences of intention applies unchanged: if differences of will are eternal together with their cause, they should all arise at once; if differences of will await further causes, one dharma is not the sole cause; "that Īśvara's will has no difference between its earlier and later stages."

Putting the entire sequence into the content of an eternal decree does not escape this entry either. Its content may describe time perfectly well. If its bearer is unaffected by what it describes, however, the world's arising at a particular moment has no corresponding event on the agent's side. This is still the price of idle spinning. For configurations accepting this further requirement, this route yields a contradiction in the temporal version and records a cost in the eternal-decree version; the two results are entered separately.

Route four: omniscience and freedom. This route considers configurations that also affirm the free choice of creatures: the omniscient being eternally knows what those creatures will do. This argument does not rely on misplaced necessity: knowing entails truth, not necessary truth. It sets out only three exhaustive branches. First, that knowledge changes over time with the creatures' choices: the knowledge undergoes change. Second, the knowledge is first mistaken and then correct: omniscience has already failed. Third, the opponent's more refined response: the knowledge neither changes nor errs; it differs only counterfactually with the creatures' choices. Had the creatures acted otherwise, that knowledge would have had different content all along. The third branch must be met directly. Counterfactual difference is a dependence in individuation. For that knowledge to be that knowledge, the identity of its content depends on what the object known does. Omniscience in this combination is an essential attribute, not an additional one. An essential attribute is individuated by something external to the being itself. "Not constituted in dependence on other dharmas" fails at the most central point. This branch has nothing to do with temporality. It concerns the first condition itself and carries more weight than the first two.

The retreat is to move knowledge outside time. Eternal knowledge does not precede choice and therefore does not lock it in. The cost is settled against the second condition. How does knowledge outside time know temporally ordered events? What is known changes, and the relation between the knower and the known changes with it. If none of this relational change falls on the knower, the knowledge's relation to the world cannot be explained. If it does, there is still change. Under the atemporal retreat, route four yields only a dilemma concerning temporality. The direct damage to the first condition is done by the third branch. Each result is recorded at its actual strength. This family of problems likewise has a precedent of direct argument within Buddhism. Fascicle 11 of the Mahāvibhāṣā provides one, exhibited in Essay 16, Chapter 2 as an instance of actual argument.² The point taken here is that the difficulty is not confined to any one religious tradition. This route concerns configurations that combine omniscience as an essential attribute of a being with self-nature and the freedom of the creatures whose actions it knows.

Route five: identity and difference between creator and creation. Chapter 3's formal structure now applies directly. If creator and creation are one, creation is self-emanation, and what is created fails to be created. This alternative is also rejected by the doctrine itself. The distinction between creating and being created is a fundamental commitment of this combination. If they are different, then by the combination's definition the creator's end is independently different, E3 and D2. All the old problems of the difference alternative arrive. The bridge of creation must have one end on a being that depends on no other dharma, so the bridge cannot be built onto it. Function must reach both ends, and reaching is dependence.

There is a finer question too, repeating Chapter 4's retrospective identification of a condition. The title "creator" is constituted through what is created; without creation there is no name "creator." A title depending on a work is standard relational individuation. Yet this candidate's identity is required to be settled prior to all relations. Either it is a creator by essence, in which case creation is as necessary as it is, and freedom is lost; or "creator" is only its temporary title, in which case its original identity needs a separate statement. Statement returns to route one's test of grammar. Chapter 3's clause fixing the scope of application fits exactly here. This dilemma has force only against E3 theorists, and this combination never conceals its E3 position.

The inventory passed on from route two is now completed. The clause denying a real relation amounts to a doctrinal admission that one end of the bridge is not really in the relation. The created being really depends on the creator, while the creator does not really stand in relation to the created being. This is a relation of dependence real at only one end: x is really related by R to y, while y is not in R. Something of this form is a new metaphysical construction. This chapter retains its legitimacy as an open issue for direct dispute, without deciding on behalf of either side. It records only the price. To preserve the first condition, creation is left suspended at one end, whereas every dynamic duty this combination assigns itself, creating and knowing the world, requires a real relation at both ends.

Three doors revisited. The retreats left open in Chapter 2 cost most in a theological setting. Here they are assessed side by side. Door one, abandon logic. Once the law of noncontradiction is withdrawn, a creed no longer excludes its opposite. The most solemn affirmation and its blasphemous denial no longer differ. The distinction between orthodoxy and heresy, and all work deriving consequences from doctrine, are scrapped at once. Every proof advanced for this candidate simultaneously fails. Besides, mature theological traditions themselves interpret omnipotence as excluding the making of contradictions. Taking this door means withdrawing from the case, not winning it. If the visitor is changing the rules rather than overturning the table, please enter door three for a separate account.

Door two, abandon content. Keep logic but give up statability, saying only what it is not. The distinction between two negations reveals this route's form: the limiting form of implicative negation, rejecting every predicate while preserving a referent. Only one step separates it from pure negation, and that step is the whole point at issue. The preserved referent is precisely the D3 assertion itself. Serious traditions have actually maintained this position. This chapter does not attack them but reads out the position's own clauses, which admit two readings. On the first, it still asserts that the referent really exists with intrinsic nature. This is a standard D3-type assertion and is refuted together with every assertion of its kind, without exemption. Record it under the strong version. On the second, it truly gives up statability, negating everything without positing anything and advancing no further assertion. Nothing remains in hand to refute. What cannot be refuted is this stance of no longer advancing a thesis, not any proposition. This is the fourth of Chapter 4's four closing routes. It passes to T-middle and the thesis domain and is recorded under the middle version.

Door three, accept true contradictions. Retain logic's normativity but change its internal rules, declaring the contradictions in this combination true without explosion. This door differs from overturning the table. It belongs among the routes of reasoning, and Chapter 7 has respected it as an opponent. The price has three entries. First, this door does not refute the derivation, only explosion. Whoever enters accepts every contradiction derived by this chapter's five routes in full. The proof part of the theorem is received in its entirety. The only remaining dispute is whether contradictions are admissible. Second, doctrinal work must be rewritten as a whole. Orthodoxy and heresy have been distinguished by consistency. A new metarule must now decide where contradiction is sacred and where it is blasphemous, and that metarule itself must not in turn be contradictory. Third, Chapter 7's textual reason for rejecting dialetheism is unavailable outside the Buddhist domain. Only the reason from cost remains, and this weighs more heavily in a doctrinal setting, since a system of doctrine relies more deeply on derivability than any treatise does. Finding: of the reasoning retreats, this door is the most expensive. Those who enter no longer dispute with this chapter whether contradictions are present, only whether they are admissible.

The combined account. Each of the five routes occupies a doctrinal joint. Their inferences come from one source, and together they show one thing: this combination of concepts gathers into itself all the features the theorem addresses. The three conditions explicitly written into its definition form the static half; creation, omniscience, and will form the dynamic half. The static half cancels identity and difference, while the dynamic half needs them everywhere. The theorem now receives its shortest formulation outside Buddhism: by definition, this combination requires every relational duty and, by definition, cancels qualification for relational work. Only one inference is used. This is no weakness but the shape of the argument. The theorem appears once at each of five joints. Their number demonstrates coverage; the single inference shows how the argument has been distilled. The predicament comes from logical structure, not faith, from the packaging, not the object believed in.

Two statements conclude. First, this chapter has used no Buddhist premise throughout. The two passages cited from the Abhidharmakośabhāṣya and the Mahāvibhāṣā are exhibits of precedent, showing that ancient thinkers made these arguments. They bear none of the weight of this chapter's proof. Independence from Buddhism has now passed from assertion to demonstration. Second, this chapter is no argument for atheism. The theorem rejects the packaging of real existence plus intrinsic nature. Forms of faith that sustain themselves through a relational reading, and practices that place themselves at the level of E2, all lie outside this chapter's scope. But this statement will not soften one cost by its tone. It says plainly: for believers who insist that their object must really exist with intrinsic nature and require that insistence to hold up in argument, the theorem is not neutral. The content of their creed is judged together with its packaging. The target of the verdict must be clear. It does not forbid faith. It forbids wanting to believe while also insisting that logic must make that faith convincing to oneself and to others. Faith as a commitment need advance no thesis. Once it does, and the thesis is again real existence plus intrinsic nature, it enters the scope of application and is handled by the rule. Such believers still have routes open: the three doors revisited above, each with its own price marked on it. The theorem settles accounts with metaphysical packaging, not with what the human heart seeks.

A restatement in plain language follows the same rule as before: change only the language, add no content, and let the first half govern any discrepancy.

The first seven chapters worked within the Buddhist household. This one takes the theorem out the door. The test case is a combination of concepts: eternal, self-existent, meaning dependent on nothing else, immutable, omniscient, and creating the world by free will. We name no existing religion and quote no scripture. We examine only this combination's own logical structure. Whoever does not use it is outside our jurisdiction. State this in advance: every conclusion in the chapter is conditional. Whoever accepts this combination bears these results. Which tradition accepts it is for that tradition to say; we name none on its behalf.

We chose it because it is a ready textbook case. When examining dharma entities and atoms, we had to dig the three conditions of intrinsic nature out of the opponent's usage. This combination writes them in the definition's first line: depending on nothing else for existence, eternally immutable, uncreated. One wording needs close attention. The definition says existence depends on nothing else. Our first condition requires identity to depend on nothing else. One clause must bridge the gap. Mature versions of this combination have always come with a simplicity clause: no internal distinction, essence being existence. Once the clause is signed, independence is complete. Versions that do not sign accept only half the independence and move outside the scope under Chapter 2's conditional formulation, at the cost of failing to live up to the name of self-existence. There is thus only one check: do they accept full independence? Mainstream versions have always done so. They then assign the candidate the heaviest relational duties: creation, omniscience, will. If the theorem works, contradictions must be thickest here.

The first route is that test of grammar. Ask the self-existent being to introduce himself. If he says he is such and such, every adjective classifies. Classification is comparison, comparison must involve something else, and his self-existence expressly forbids this. So grammar leaves him only one sentence: "I am I." Be clear about the claim. We use no scripture as evidence. This is a theorem of grammar stated in advance: a self-existent being who introduces himself on an ontological reading can say only this. Whether there is such a sentence in the world and how it should be read belong to philology. Sentences of this kind have always admitted two readings. Read as a promise to be with those addressed, it lives entirely at the ordinary level, mutually dependent, responsive to conditions, and functional. Our theorem does not touch a word of it. Read as an ontological declaration, "the self-subsisting being," "existence itself," it enters Chapter 4's dilemma exactly. The theorem supplies only a sorting format: the promise reading is unharmed; the ontological reading enters the dilemma. It says nothing about which sentence belongs to which reading.

The second route asks how an unchanging agent can act. Creation is an action, and action has a before and after. What is eternally immutable has no before and after. If there is no difference on his side between before he acts and the moment he acts, this cannot count as his doing anything. At most the world has appeared on its own. If there is a difference on his side, he has changed. The Sarvāstivāda's dharma entities and their functions faced this same problem. Essay 15 went through it once. Here it runs again with a different object, without changing a word. Will is the same. If an eternal will eternally wants to create, the world should be as old as he is. If the world has a beginning, the will has turned over at that beginning, and turning over is change. The opponent's strongest version says that the relation of creating is real only on the world's side. There is no relation at all on his side, hence no change to discuss. This step does evade contradiction, and the cost is recorded on the spot. Creating or not creating makes no real difference on his side, and the title "creator" is supported entirely from the world's side. This entry passes to the fifth route for assessment.

Third, the problem of the first moment. This route adds a condition: it addresses those who also affirm complete goodness and require a reason why the world begins at this time and in this form. First speak precisely. Creating by free will does not mean that everything has only one cause. Cause and effect within the world can perfectly well arrange its sequence. The trouble is all at the first moment. There is no world yet, and only the eternal being himself is available to rely on. If the cause is always ready and always at work, all the effects should emerge at once. The opponent naturally says it is not always at work. But there must be some reason for when it works and when it does not. A reason in external conditions? At the first moment there is no outside. A reason in his own change? He does not change. The ancient questions put to the sole-cause theorist apply here unchanged. Doing different things at different times requires different thoughts at different times. If the thoughts are as eternal as he is, they should occur together. If they must await other conditions, he is not the sole source.

Writing the entire schedule into an eternal decree makes no difference. Time being written into a decree does not mean the decree itself is moving. The world comes into being at that moment, but nothing corresponding happens on his side. This is still idle spinning. Under that added condition, this route yields a contradiction for a creator within time and records a cost for the eternal-decree version; the two results are kept separate.

Fourth, can omniscience coexist with free choice? Here the case is knowledge of creatures who are also held to choose freely. He eternally knows what a creature will do on Tuesday. Can that creature still choose otherwise on Tuesday? If it can, either the knowledge changes with its choice, in which case he changes, or the earlier knowledge was wrong, in which case omniscience is lost. The opponent has a third and more refined response: the knowledge neither changes nor errs; it differs only counterfactually. Had the creature acted otherwise, that eternal knowledge would have had different content all along. Meet this directly. If the knowledge's content follows the creature's actions, the knowledge's identity is determined by someone external. Omniscience is his essence, not something added. If his essence is determined from outside, independence fails at the most important point. This branch does not concern time. It targets the first condition and weighs more heavily than the first two.

Someone moves knowledge outside time. Eternal knowledge does not precede choice and cannot lock it in. The cost is entered against immutability. How does knowledge outside time know things in time? The world changes, so the relation between knowledge and the world changes. If none of that change falls on the knower, the knowledge has no relation to the world. If it does, he changes. Problems of this kind have been debated directly within Buddhism too. The Mahāvibhāṣā has a ready example, exhibited in Essay 16. The difficulty does not belong to any one religious tradition: this route examines an intrinsically independent being's essential omniscience together with the free actions of its creatures.

The fifth route asks whether creator and creation are one or two. If one, there is no creation, at most self-emanation, and the doctrine itself forbids this route. If two, by definition, the creator is one of the pair and is independently established. All Chapter 3's old problems arrive together. The bridge must be built onto the end that depends on nothing, so it cannot be built. Function must reach both ends, and reaching is dependence. A finer point remains. The title "creator" is bestowed only once there is creation, just as a condition is retrospectively identified by its effect. The title depends on the work, yet the candidate's identity must be settled before all relations. If he is a creator by essence, creation is as necessary as he is, and freedom is gone. If the title is temporary, what is his original identity? That returns to the first route's "I am I."

The entry transferred from the second route is assessed here. Saying that the relation is real only on the world's side admits in black and white that the bridge's other end is not attached to him. The world really depends on him, yet he has no real relation to it: a dependence real at only one end. We do not judge on anyone's behalf whether such a new construction holds. We record only the price. To preserve independence, creation is left suspended on one side, yet not one of the jobs this combination assigns itself, making and knowing the world, can do without reality at both ends.

Look at the three doors in turn. Overturn the table and refuse logic: a creed no longer excludes its opposite. Its most solemn affirmation cannot be distinguished from words that blaspheme it. Orthodoxy and heresy cannot be distinguished, and its own proofs all fail. Seasoned theologians themselves say that omnipotence does not include making contradictions. What this door buys is forfeiture, not victory. Anyone arriving to change the rules, rather than overturn the table, should take the third door.

The second door gives up content, saying only what he is not. The two kinds of negation expose this route's form. Negating to the end while preserving an unsayable object stops just one step short of negating all the way, and that step is the whole dispute. Serious traditions stand on this route. We do not press an attack but read their own clauses aloud. These have two readings. Those who still insist that the unsayable object exists solidly have this assertion refuted alongside other assertions of its kind, without exemption. Those who truly negate without positing and utter no claim are left holding nothing to refute. What cannot be refuted is the stance of no longer advancing a thesis, not any sentence. The latter reading is recorded under the middle version.

The third door accepts true contradictions. Keep the table, change the rules, and say these contradictions are all true without explosion. This route reasons, and Chapter 7 respected it. The price has three entries. Entering this door means accepting every contradiction our five routes derive in full. The case concerns only whether contradictions are admissible. All doctrinal housekeeping must be redone. Orthodoxy and heresy were distinguished by consistency; now a new rule must decide which contradictions are sacred and which blasphemous, and the rule itself must not in turn be contradictory. Chapter 7's scriptural reason for rejecting this school cannot be carried outside the door. Only the cost reason remains, and doctrine relies even more heavily on derivability than a treatise. So this is the costliest reasoning retreat. Those who enter no longer dispute with us whether contradictions are present, only whether they are admissible.

Taken together, half this combination's definition cancels identity and difference: self-existence, immutability, being uncreated. The other half, its duties of creation, omniscience, and will, needs identity and difference everywhere. The theorem's version beyond Buddhism is one sentence: by definition, all relational jobs are required; by definition, qualification to do them is cancelled. The predicament is in the packaging, not the faith.

Two final points must be clear. This chapter uses not one Buddhist premise. The two quotations from the Abhidharmakośabhāṣya and the Mahāvibhāṣā exhibit ancient thinkers making the same arguments; they do no work in the proof. Nor is this chapter a proof of atheism. What is refuted is the packaging of solid reality plus being established on its own. Faith read through relations and promises, and practice placed at the ordinary level, are outside the scope. One cost must be stated plainly, without talking around it. Whoever insists that the being believed in must be solidly real and established on its own, and requires that insistence to make rational sense, cannot find the theorem neutral. The creed is judged along with the packaging. Be clear about the judgement. It does not forbid belief. It forbids wanting to believe while also demanding that logic convince oneself and others. Keep faith without advancing a thesis, and no one touches you. Speak a thesis, and if it is solid reality plus being established on its own, it enters the scope and is handled by the rule. Such a person still has routes available. The three doors above remain open, with a price marked on each. The theorem settles accounts with packaging, not the human heart.

References¶