Details of Ways of Knowing: 14 Number of Valid Ways of Knowing
Extensive Explanation of “Compendium of Ways of Knowing”
According to the Buddhist tenet systems in general, there are only two valid ways of knowing: bare cognition and inferential cognition. Within the various systems of Buddhist tenets, the definitions, divisions and constituents of each may differ, but all agree that the count of distinctive types definitely comes to only two. To conceive of less than this amount is a distorted conceptual cognition that repudiates what is existent, while to conceive of more is one that interpolates something extra that does not accord with fact.
Various non-Buddhist Indian schools assert such repudiating and interpolating theories concerning the count of valid ways of knowing. But just as in the Buddhist fold different schools of tenets have different definitions, likewise the various non-Buddhist systems have their own individual presentations for what a valid way of knowing, bare cognition, inferential cognition and so forth are, what objects can be cognized by each and so on. Each of these systems paints its own comprehensive and coherent picture of how the mind works and how you can become aware of all knowable phenomena. It is beyond the scope of the present work, however, to explain each of the following non-Buddhist schools and their individual definitions in their own terms.
What is at issue here is not the definitions asserted by each of these schools, nor is this the place to point out and debate internal logical inconsistencies in their theories. That is a separate subject matter altogether, treated in a different class of study. Therefore, the intended purpose here is not to consider the following non-Buddhist assertions within the context of their own basic definitions. The only concern is regarding the number of distinct valid ways of knowing there are. Thus, the number asserted by each school is considered on its own and, moreover, within the context of the Buddhist Sautrantika definitions of valid ways of knowing, bare cognition, inferential cognition and what can be cognized by each.
By examining these non-Buddhist assertions and considering their implications if accepted within the Sautrantika context, you will gain conviction in the logical consistency of this Buddhist system and understand why only two valid ways of knowing need to be posited. This will then safeguard you from the dangers of reading about a non-Buddhist system and borrowing certain theories from it that either repudiate Buddha’s words or interpolate something extra. To do so adulterates the teachings and creates serious obstacles to gaining realizations through its paths. The best way to prevent this is to understand how such deletions or additions are either totally unnecessary or how they create other problems by generating internal inconsistencies or absurd conclusions. All Buddhist and non-Buddhist philosophies present pathways to their various goals, even though these final goals may differ. If you mix the teachings of one system with another, however, no achievement of any goal is possible, and it weakens both systems through confusion and adulteration.
Only One Valid Way of Knowing
The Charvaka (rgyang-’phen-pa) school of materialists accept only bare cognition as a valid means of cognition because, according to its system, only obvious phenomena exist. Logical pervasions, being non-material and obscure, do not exist.
According to the Sautrantika theories, as a limited being you can explicitly cognize only obvious phenomena through bare cognition, but not anything obscure. To cognize obscure phenomena explicitly, you must rely on valid lines of reasoning with inferential cognition.
Something obvious, then, is defined as that which can be explicitly apprehended by valid bare cognition. Obvious objects are synonymous with all functional phenomena, in other words, all nonstatic objective entities.
Something obscure, on the other hand, is defined as that which can be explicitly apprehended by valid inferential cognition. This is synonymous with all validly knowable phenomena, both nonstatic and static.
Obscure phenomena, then, may be static or nonstatic, and although the static ones, being metaphysical entities such as absences and categories, are only conventionally true phenomena and lack truly established existence (rang-bzhin-gyis grub-pa) since they do not perform any functions, nevertheless they are existent. Their existence is established by their self-natures (rang-bzhin-gyis grub-pa) and individual defining characteristics (rang-gi mtshan-nyis-gyis grub-pa) found on their own sides. Therefore, they can be validly cognized. Absences, such as the absence of an apple on a table, can be validly cognized implicitly by valid bare cognition. Categories, on the other hand, can only be cognized validly by inferential cognition by relying on a valid line of reasoning.
All metaphysical entities, however, can be apprehended by a Buddha implicitly with bare cognition. A Buddha need not rely on inferential cognition to cognize those that are categories. Furthermore, obscure nonstatic phenomena, such as subtle nonstaticness, that can only be known by limited beings by reliance on a line of reasoning are obvious to a Buddha and known explicitly by his bare cognition. Therefore, it could be argued that valid bare cognition is the only valid way of knowing because only it is non-conceptual and therefore non-deceptive.
Such a proposition, however, has several absurd conclusions. If it were true, then only a Buddha could validly know something obscure to an unenlightened limited being’s sensory bare cognition, for instance the presence of fire on a smoky mountain, or the presence of someone behind the house whom you cannot explicitly see but whom you can hear. But this contradicts worldly experience.
Although limited beings might not be able to see a fire on a distant smoky mountain, yet when they see the smoke, they know there must be a forest fire present, for instance, and begin to take steps to extinguish it. Fire itself is a functional, objective phenomenon and they cannot explicitly see it. But what about the presence of fire at a smoky place? How could they have known of it?
If the presence of fire there could have been known only by bare cognition, it must have been either explicitly or implicitly apprehended by it. If the former were the case, it would need to have been an appearing object to that person’s visual consciousness, for instance. But it did not appear, since it was obscure. The only other possibility, then, is that the presence or existence of fire there had been implicitly apprehended by this visual bare cognition.
At this point in the argument, it should be recalled that only objective entities are explicitly apprehensible by bare cognition and only certain metaphysical ones, such as the absence of fire or the presence of fire, are implicitly apprehensible, according to the Sautrantika theories.
If the presence of fire had been implicitly apprehended, then the presence of fire would have to have been an imputedly existent and imputedly knowable phenomenon. What it establishes as present is fire and, as an imputation phenomenon, it must have a basis for imputation, and that would be the place where smoke is present. To apprehend the presence of fire there, this basis would need to appear first and then be apprehended simultaneously with its basis. Both of these assertions are true.
The presence of something, although a metaphysical entity, is a validly knowable phenomenon. As such, its existence as a validly knowable phenomenon is established by its individual self-establishing nature and individual defining characteristic findable on its own side. But because it does not perform a function and therefore its existence as a validly knowable phenomenon cannot be truly established by that criterion – it cannot be established by seeing it do something – its existence as a validly knowable phenomenon can only be established in terms of the words and concept for it. The existence of the presence of fire, then, is imputedly established as being what the words and concept for it refer to as mentally labeled or applied to the something that is its basis – in this case, a place where smoke is present.
Since the existence of the presence of fire at a place where smoke is present can only be established dependently on the mechanism of mental labeling, then for it to be a valid mental labeling, the labeling must rely dependently on not only valid observation, but also, when it cannot be observed, on a valid line of reasoning. Such a line of reasoning would be: “A place where smoke exists is a place where fire exists, because there is smoke there, and fire is the cause of smoke.”
Thus, even if it is granted that the existence and presence of fire, even on a distant smokey mountainside, might be implicitly apprehensible by valid sensory bare cognition, it would have to have been validly cognized first by an inferential cognition that relied on the pervasion of fire and “being a cause of smoke.” Therefore, there must be more than one valid way of validly knowing such an obscure phenomenon as the presence of fire where there is the presence of smoke, and thus there is not just one valid way of knowing as the Charvakas claim.
Furthermore, mentally challenged persons and small children need to be taught such facts as where there is smoke there is fire. They do not automatically apprehend the presence of fire when they see a distant smoke-filled place. Therefore, since this is the case, most people will know such obscure phenomena primarily by relying on a valid line of reasoning and inferential cognition. Otherwise, most people would remain totally ignorant of anything obscure.
Next consider how you know of the presence of a man standing behind a house. By relying on the information of someone reliable standing to the side of the house who actually does see him, you can know of the presence of the man behind the house even though you yourself do not have bare cognition of him. Although the person from whom you gained this information saw the man in question by sensory bare cognition himself, you did not. If you did not rely on an inferential process with the line of reasoning: “What this
person who has sensory bare cognition of the man behind the house says about his presence there is true because he has sensory bare cognition of him,” how else do you know of this man’s presence yourself? Otherwise, if you did not explicitly see something yourself, you would not be able to know it and could never gain definite knowledge of it by relying on someone else as a valid source of information.
Of course, you could object and say that the person giving you the information might be lying. That raises the issue of how to know that someone is a valid source of information. To ascertain that, you would need to rely on reasons for considering them a valid source of information, which means you need to rely on inference.
Suppose you broke the bone in your leg. Even though you might see signs of this on the surface of your skin, you could never know that your bone was broken unless you saw the bone itself. You could not infer it from the surface signs, nor could you believe a physician who so inferred if you asserted that only bare cognition is a valid way of knowing. Even if you saw an x-ray picture of the broken bone, this would not do since you would have to rely on the line of reasoning: “What I see in this picture is my actual bone because it is a valid x-ray of it.” Otherwise, why should you believe that seeing this picture serves as a way of seeing your bone? Someone who did not know of the existence of the x-ray process would certainly not believe this picture.
Therefore, because such absurd conclusions follow, it is a distorted cognition to accept, as the Charvakas do, only one valid way of knowing, namely bare cognition, and to repudiate the validity of inferential cognition.
Valid cognition through bare cognition (Skt. pratyakshapramana, Tib. mngon-sum tshad-ma). This is cognition of what is lucid or clear (Skt. vishada, Tib. dvangs-ba), in other words, what Sautrantika calls obvious phenomena. It does not require any other cognition in order to arise.
Valid cognition through cognition of what is beyond observation (Skt. parokshapramana, Tib. lkog-gyur tshad-ma). The term “beyond observation” (Skt. paroksha, Tib. lkog-gyur-ba) is the same term Sautrantika uses for obscure phenomena. It literally means “beyond the eyes.” It requires reliance on a prior bare cognition in order to arise.
Our text includes the Jains with the Charvakas as asserting only bare cognition as the single valid way of knowing. We could speculate that this is because, since valid cognition of obscure phenomena is derivative of valid cognition of obvious phenomena, ultimately all valid cognition comes down to bare cognition of what is obvious, namely observable phenomena. But, since that is purely a speculative guess, let’s analyze the issue more deeply.
Inferential cognition based on the force of the actuality of phenomena (dngos-stobs rjes-dpag) relies on logical pervasions derived from observing causal relations (Skt. tadutpatti, “arising from that”) or equivalent identity-natures (Skt. tadatmya, “being in the same identity-nature as that”).
Inferential cognition based on what is well-known (grags-pa’i rjes-dpag) relies on having heard and learned conventions. Inferential cognition based on conviction (yid-ches rjes-dpag) relies on having heard or read valid sources of information. Nevertheless, at the time of a valid inferential cognition, the cognition derives from direct reliance on a valid line of reasoning; it does not directly rely on bare cognition of something obvious in the immediately preceding moment in order to arise. It is not
like the cognition of imputation phenomena, such as the perishing and gross impermanence of a clay jug when it falls off the table and breaks into pieces. Cognition of these imputation phenomena requires first bare cognition of the clay jug as it falls and breaks, immediately followed by bare cognition of the clay jug, its perishing and its gross impermanence all together. The perishing and the gross impermanence of the clay jug, however, are obvious phenomena, not obscure. They explicitly appear to bare cognition.
Since valid inferential cognition of an obscure object does not arise in the way in which valid bare cognition of an imputation phenomenon arises and is not itself a bare cognition, it would not fall into either of the two divisions of valid ways of knowing posited by the Jains.
Conceptual schemes for dividing all validly knowing phenomena may include:
Categories or sets that have a common locus (gzhi-mthun) of items that are members of more than one set, such as the sets of obvious phenomena and obscure phenomena. The set of obvious phenomena includes only those nonstatic phenomena that can be explicitly apprehended by bare cognition. The set of obscure phenomena includes all validly knowable phenomena. In this division scheme of obvious and obscure phenomena, nonstatic phenomena that can be explicitly apprehended by bare cognition are members of both sets of phenomena.
Categories or sets that are mutually exclusive (‘gal-ba), having no items that constitute a common locus that are members of both sets. For example, non-static phenomena and static phenomena. There is no validly knowable phenomenon that is a member of both sets. Conceptual schemes dividing all ways of knowing in general, however, only include sets that have a common locus:
However, for conceptual schemes dividing only valid ways of knowing, there is only the mutually exclusive sets of valid bare cognition and valid inferential cognition. There is no valid way of knowing that is a member of both sets.
Similar to the Jain division scheme of valid ways of knowing, Sautrantika differentiates between cognition of self-sufficiently knowable phenomena (rdzas-yod) and cognition of imputedly knowable phenomena (btags-yod). But these two sets of ways of knowing are not mutually exclusive, because valid cognition and non-valid cognition may be members of both sets.
Even when the division is specified in terms of valid cognition of self-sufficiently knowable phenomena and valid cognition of imputedly knowable phenomena, they still do not constitute mutually exclusive sets of valid ways of knowing. Both valid bare cognition and valid inferential cognition may validly cognize both self-sufficiently knowable phenomena (such as sound when hearing it and sound as being subtly impermanent) and imputably knowable phenomena (such as a person when seeing someone and a person as neither one nor many with the five aggregates). Therefore, cognition of self-sufficiently knowable phenomena and cognition of imputedly knowable phenomena do not constitute mutually exclusive sets as a way to divide all valid ways of knowing.
Therefore, the Jain scheme of dividing valid ways of knowing into valid cognition through bare cognition and valid cognition through cognition of what is beyond observation actually is just a scheme having one valid way of knowing with these two divisions constituting a non-dichotomous pair as subsets. In order to specify distinct, mutually exclusive valid ways of knowing having no common locus, it is necessary to specify two ways.
A mutually exclusive pair that, although lacking any member that is both, nevertheless does not preclude other valid ways of knowing that are members of neither of the two.
Three Valid Ways of Knowing
When you understand correctly and decisively what someone means by hearing the sound of what they say, or you learn something that is true by reading it in a text of scriptural authority or hearing it explained by someone trustworthy, you have validly known something through a verbal indication (sgra-byung tshad-ma, Skt. shabdapramana). The Sautrantika position would agree that such knowings are valid; this is not in dispute. The point of contention is whether or not these constitute a third separate category of valid ways of knowing.
If there were a third way, then the above types of knowing something through verbal indication would have to be gained by a process that involved neither bare cognition nor inferential cognition. But this is absurd. When you know what someone means by hearing their speech, surely the first step involved is valid auditory bare cognition of the sound. Furthermore, there are certain popular conventions that assign meanings to specific combinations of audible phonemes within one language. If you have prior knowledge of
these conventions, plus you know the audio categories of the sounds of the words and the meaning/object categories of what they mean, then when you hear certain sounds you know what they mean through application of these categories. If this conceptual process does not rely on a line of reasoning such as: “This combination of phonemes has such and such a meaning because that is the popular convention,” then how else does it work? Clearly the process of such a way of knowing is one of inferential cognition following an initial bare cognition of the sound. It is a distorted conceptual cognition to conceive of this way of knowing as functioning independently and separately from these two.
The same type of analysis can be used to understand how you gain valid knowledge from reading a text of scriptural authority. First you must have visual bare cognition of a written representation of certain words, which also have phonemic elements associated with them by popular convention. Next, by having prior knowledge of the conventions with which the words are associated with the meanings they represent, you understand the meaning of the words when you read them. This occurs through a process of inferential cognition as
explained above. Finally, you know that what you are reading is correct information with scriptural authority by relying on such lines of reasoning as: “The information in this book is correct because it was written by a valid and reliable author” and “the author of this book is a valid source of information because of whatever reason it is that you are confident that they are so.” Thus, the final step involved in this way of knowing is also an inferential cognition.
Another example is your friend is in the next room and you are having a conversation with her. You hear the sound of her voice but cannot see her. The Samkhya school would explain that you know of her presence in the next room by verbal indication. They argue that your friend’s presence is not explicitly heard by your auditory bare cognition and you do not separately infer it. Therefore, such knowledge must be gained by a valid way of knowing other then bare cognition or inferential cognition. But is it necessarily the case that a third valid way of knowing must be posited?
It is true that your friend’s presence in the next room is not obvious to your auditory bare cognition and is not its appearing object. Her presence is obscure. It can be, however, inferred by the line of reasoning: “My friend is present at the place from where I hear her voice in conversation with me because the sound of her responses to my words is being produced.” Having a conversation with her and this line of reasoning excludes the possibility that the sound of her voice is coming from a recording device, like a voice message you hear on your phone.
But you could argue that you are not explicitly inferring that while hearing her voice. You simply know she is there. Well then, how do you know this? According to the Sautrantika theories, you can also have such knowledge by implicit apprehension at the same time as explicitly apprehending the sound of your friend’s voice with auditory bare cognition.
Her presence is an imputation phenomenon on the basis of the sound of her voice and can be confirmed by an inferential cognition relying on the above line of reasoning. Thus, it is possible to know of her presence either implicitly by auditory bare cognition or explicitly by inferential cognition.
In the case of seeing smoke and implicitly apprehending the presence of fire as an imputation phenomenon on it as its basis, the point of the discussion was totally different. There the assertion refuted was that the presence of fire could only be known by valid bare cognition and not by inferential cognition. Here, although the presence of your friend when you hear her voice coming from the next room is likewise obscure, as is the presence of fire when seeing smoke, the assertion refuted is that you know it by a third process of valid knowing, which is neither bare cognition nor inferential cognition. As has been shown, it is either known implicitly by bare cognition or explicitly by inference, but not by a third process that is neither of these two.
Four Valid Ways of Knowing
The Nyayas accept four, adding to these three (4) comprehension of an analogous example. You know what a brahmin cow is: it is bovine animal with a fleshly hump on its shoulders and a dewlap hanging from its neck. You see a water-buffalo. To comprehend that this is like a brahmin cow but without a fleshly hump on its shoulders and a dewlap hanging from its neck is to validly cognize something through comprehension of an analogous example (dpe-nyer ‘jal-ba, Skt. upamana). You do not know what exactly the animal is, but by the analogy of what a brahmin cow is, you know one thing that it is not. It is not a brahmin cow.
Another example is looking for your friend in a crowd, seeing a stranger and knowing by the analogy of your friend that this person is not the one you are looking for. But is knowing something through comprehension of an analogous example a separate valid way of knowing different from bare cognition or inferential cognition?
Previously, the type of conceptual cognition that applies a name has been explained. To be able to conceive of a bovine animal with a fleshy hump and dewlap as a brahmin cow depends explicitly on the line of reasoning: “A bovine animal with a fleshy hump and dewlap is a brahmin cow because that is the popular convention of how such a brahmin cow is defined.” Therefore to identify what the animal
is, you rely on a process of inferential cognition. The same is true when identifying what a certain animal is not. Surely when knowing that a water-buffalo is not a brahmin cow, you can come to this conclusion by relying on the line of reasoning: “This bovine animal without a fleshy hump and dewlap is not a brahmin cow because a brahmin cow has such a hump and dewlap and this animal has neither.” This is a form of inferential cognition.
You could also know that this water-buffalo is not a brahmin cow through visual bare cognition. It is similar to seeing a bare table-top and also seeing that absence of the vase on it. In explicitly apprehending the former, you can implicitly apprehend the latter absence that is an imputation phenomenon on the table as its basis. Likewise, in explicitly apprehending the water-buffalo with
visual bare cognition, although your sensory consciousness as an ordinary being cannot identify the animal as a water-buffalo, it can still implicitly apprehend the absence of a brahmin cow as an imputation phenomenon on it as its basis. To be able to apprehend implicitly either the absence of a vase on a bare table-top or the absence of a brahmin cow on the basis of a water-buffalo, you need previously to have been acquainted with either the vase or a brahmin cow. But at the time of explicitly seeing the table-top or the water-buffalo, you can implicitly notice absences without reliance at that time on a line of reasoning.
Thus, it is possible to validly cognize something through apprehension of an analogous example through either implicit bare cognition or inferential cognition. There is no need to posit it as a separate way of knowing that is neither of these two. To do so is a distorted conceptual cognition that interpolates something extra and unnecessary.
There is one further instance that can be cited as valid cognition through comprehension of an analogous example. This would be, for instance, cognizing the impermanence of sound simply from cognizing; “Sound is impermanent, like a clay jug.” But this is an incomplete thesis. The full thesis would be: “Sound is impermanent because it is an affected phenomenon, like a clay jug.” But it could only be validly cognized by inferential cognition by relying on the three components of the line of reasoning. Even to cognize the full thesis by itself, without reliance on the line of reasoning, let alone cognizing the impermanence of sound simply by relying on the analogous example of a clay jug, would be merely a presumptive cognition, not a valid cognition.
Six Valid Ways of Knowing
The Bhatta Mimamsakas are followers of the writings of the philosopher Kumarila Bhatta, who lived around the turn of the eighth century CE. The slightly earlier Mimamsaka philosopher, Prabhakara, had added valid cognition through implication to the four valid ways of knowing that Nyaya assert. Kumarila Bhatta further added valid cognition of non-existence to Prabhakara’s list of five. Both the Prabhakara and Bhatta branches of Mimamsaka are called Purva Mimamsa (Early Mimamsa) since they are followers of the earlier Upanishads; whereas the Vedanta schools that later followed, known as Uttara Mimamsa (Later Mimamsa), follow the later Upanishads.
 The fat man Devadatta does not eat during the day. Because Devadatta is fat and because people must eat in order to be fat and can do so during either the day or the night, you know by implication (don-gyis go-ba, Skt. arthapatti) or disjunctive reasoning that fat Devadatta must eat at night. Since such knowledge of something obscure is gained by reliance on a valid line of reasoning, such as: “To be fat requires eating; eating can be done either during the day or the night; fat Devadatta does not eat during the day; therefore fat Devadatta must eat at night, because he is fat,” there is no need to specify a separate valid way of knowing for it. Understanding by implication, then, can be included in the category of valid inferential cognition.
Another example is you know someone is in your two-room house, but you do not see them in the front room. By implication or a process of elimination you know they must be in the back one when you implicitly apprehend their absence in the front room upon visual bare cognition of the empty room.
 There are four types of nonexistence (dngos-po med-pa, Skt. abhava):
[6a] An antecedent nonexistence, or antecedent absence, is an absence of something not yet happening before it is presently happening. When milk is presently happening, there is the antecedent nonexistence (antecedent absence) of not-yet-happening yogurt that is also presently happening as an imputation phenomenon on the basis of the presently-happening milk as its obtaining cause (nyer-len-gyi rgyu). An obtaining cause is the substance that will give rise to the result and, in doing so, will cease to exist, like a seed for a sprout. The absence of the not-yet-happening yogurt endures as this imputation phenomenon for so long as the presently-happening milk does not curdle.
Such an absence can be known either by implicit apprehension when explicitly apprehending the milk with visual bare cognition, or it can be inferred. In the latter case, you could infer this antecedent absence by reliance on such lines of reasoning as: “The yogurt is antecedently absent in the milk because the milk has not yet curdled; when milk curdles it becomes yogurt, the milk has not yet curdled, therefore the yogurt is absent antecedent to its production,” and so forth.
[6b] Once the milk has curdled into yogurt, you can know of the perished nonexistence (the perished absence) of the no-longer-happening milk as an imputation phenomenon on the presently-happening yogurt as its basis. Such an absence can also be known either implicitly through bare cognition or explicitly through inferential cognition similar to what has been explained with regard to antecedent absences.
[6c] To know a mutual nonexistence (a mutual absence) is to know that two sets of items have no common locus; they constitute a mutually exclusive pair with nothing that is an example of both. For instance, being a horse and being a cow are mutually exclusive. If something is one, it cannot be the other and there is nothing that can be both. When you see a horse with visual bare cognition, then, you can validly know that it could not possibly also be a cow.
But you can know this mutual absence of being a cow as an imputation phenomenon on the basis of the horse simply by implicit apprehension with this visual bare cognition. No separate valid means of knowing need be posited. You can also know such mutual absences by inferential cognition depending on various lines of reasoning, such as: “A horse is devoid of being a cow because it is a horse and there is no common locus of being a horse and being a cow.” This also requires knowledge of the definitions of the two types of animals.
[6d] When you see a rabbit’s head and know of the absolute nonexistence (absolute absence) of rabbit-horns on it, you know of the absolute absence of something that does not exist. But this too can be known either by implicit apprehension when having visual bare cognition of the rabbit’s head, or by an inferential cognition relying on such lines of reasoning as: “There are no rabbit-horns on a rabbit’s head because rabbit-horns are nonexistent; nonexistent phenomena cannot be located anywhere, because they do not exist,” and so forth.
Eleven Valid Ways of Knowing
The followers of the Charaka (Samhita) claim that the number is definitely eleven. To the above six they add valid cognition through (7) conjunctive reasoning, (8) non-perception, (9) tradition, (10) inclusion and (11) intuition.
The Charaka Samhita (Charaka Compilation), dating from perhaps 200 BCE or earlier, is one of the two main textual sources of the Ayurvedic tradition of Indian medicine. In addition to the six valid ways of knowing asserted by the Bhatta Mimamsakas, this text explained six further ways in the context of methods for diagnosis of disease.
 With inferential cognition, you use analytic reasoning to infer the cause from an effect, for instance where there is smoke there must be or must have been a fire as its cause. Such a fact is obscure, because the fire does not appear to your visual bare cognition explicitly apprehending the smoke. You must rely on a valid line of reasoning to infer the smoke’s cause, which at present may be totally nonexistent. But what about validly knowing by conjunctive reasoning (rigs-pa’i tshad-ma, Skt. yuktipramana) what effects come from the conjunction of what causes, for instance fire and burning fuel produce smoke or “if you play with matches, you will get burned”? Is this a separate category of valid ways of knowing?
Effects that come from causes may be either obvious or obscure. In the case of fire, if you watch one, you can know by valid bare cognition the production of smoke as its effect. By experiencing the sensation of a burn with your tactile bare cognition, you can also come to know the effect of playing with matches. Later, when you merely see a fire or a child playing with matches, you can recall your experience and validly infer what the result will be. A separate means for knowing such things is totally unnecessary.
There are certain effects that are either obscure or extremely obscure. For instance, what will be the long-range effect of certain political policies undertaken now? If you engage in war, what will the effect be? For a start, you can know by conjunctive reasoning, or synthetic reasoning, from previous examples of war that many people will be killed. But such knowledge comes from inferential cognition based on recollection of precedent.
As for knowing who the victor will be or what will be the type of rebirth gained by the instigators of the war, such effects are much more obscure. You may conceptually hope for a certain outcome to occur and your prediction may turn out to accord with fact, but as an ordinary being, your predictions, even if based on what seems reasonable, will not be reliable and infallible. Only a Buddha can know the results of all actions and this he does by yogic bare cognition. He knows explicitly that if you commit a destructive action, suffering will result, and if you take others’ lives, you will be reborn in a lower realm. The former type of result can even
be seen in this lifetime, but you may not realize at all the connection between the previous destructive act and the suffering result. The latter type, however, is extremely obscure because you do not ordinarily see future rebirths. But by relying on an inferential cognition based on the line of reasoning: “What a Buddha says is true because it has been spoken by a valid person” and so forth, you can know with confidence that suffering and lower rebirths result from committing destructive actions. Thus, all types of validly knowing something reasonable through conjunctive reasoning come under the category of either valid bare cognition or valid inferential cognition.
 If you do not perceive something when if it were there you would, then you know by non-perception (mi-dmigs-pa’i tshad-ma, Skt. anupalabdi) that it is not there. For instance, you can know of the absence of rabbit-horns on a rabbit’s head by your non-perception of them, because if they were there, you would surely see them. This is different from simply knowing the absolute absence of rabbit-horns, in which you know something because of the absence of an object. Here you know something because of the absence of a valid means of cognizing it.
But surely this can be implicitly apprehended by the bare cognition of your reflexive awareness explicitly apprehending the valid visual bare cognition of the rabbit’s head. It could also implicitly apprehend the absence of a valid cognition of rabbit-horns. Such information can also be gained by inferential cognition relying on such lines of reasoning as: “Rabbit-horns are absent on the rabbit’s head because, if they were present, I would surely perceive them.”
 When you validly know something by tradition (zhes-grags-pa’i tshad-ma, Skt. aitihyapramana), you believe something to be true because, literally, “it has been said to be so.” An example is knowing that a certain tree contains a spirit because all your ancestors and everyone in your community believe it does. What you know by such a means may or may not accord with fact. You may also
believe it to be true in a decisive way, based on conviction in the validity of the source of information. Or your knowledge may be indecisive if it is based on uncritical faith or fear. Thus, knowing something by tradition may be a presumption or a distorted conceptual cognition. But in the cases in which it is a valid knowing, it is an inferential cognition in which you know a piece of obscure information to be true by relying on the line of reasoning, “It is true because its source is valid.”
 When you validly know something by inclusion (srid-pa’i tshad-ma, Skt. sambhavapramana), you know about the individuals included within a group by knowing about the group itself. An example is knowing that there are at least 10 people present in the room when you know for sure there are 50. Another example is knowing that a specific piece of fruit in a basket is an orange when you know that the basket is a basket of oranges. But surely such information is known inferentially by reliance on lines of reasoning such as: “This is an orange because it is a member of a group of oranges.”
 Finally, if for no apparent reason you have an intuitional feeling or adventitious thought that your mother will visit you today and she actually does, then you knew she was coming by intuition (snyam-sems-pa’i tshad-ma). In most cases, this is a form of conceptual cognition that hopes for something to happen. If it accords with fact, it is by mere coincidence. But it is possible to
have extrasensory bare cognition of such future events. This can be gained as the result of effort in this lifetime to achieve such abilities through the attainment of shamatha and a more refined state of consciousness of one of the upper planes. You may also have such an instinctive ability from birth as the result of having achieved such absorbed concentration in a previous lifetime. In either case, however, if such intuitional knowledge is valid, it is a bare extrasensory mental cognition. There is no need for it to be asserted as a separate valid means of knowing.
The Sautrantika System of Only Two Types of Valid Cognition
Our own tradition is that it is definite that there are only two: (1) bare cognition and (2) inferential cognition. As demonstrated by the above discussion, there are many faults in positing more or less than these two valid ways of knowing, and thus it is a distorted conceptual cognition to so conceive. Therefore, according to the Buddhist traditions in general, there are only two valid ways. The reason for this, in terms of the Sautrantika definitions and assertions, can be summarized in two points:
The first point is that all existent phenomena are either obvious or obscure. The former, referring to all functional, nonstatic phenomena, can be explicitly apprehended by valid bare cognition. But although all nonstatic phenomena can be so apprehended, it does not follow that they are apprehended in this manner in every situation or by every living being. In certain situations, particular
nonstatic phenomena are obscure, and in such cases, they cannot be apprehended by the bare cognition of ordinary beings. Obscure phenomena, then, refer to all knowable phenomena, both nonstatic and static, and they can be explicitly apprehended by valid inferential cognition. Thus, these two valid ways of knowing are sufficient for knowing all existent phenomena, whether they be obvious or obscure. Any third valid way of knowing is superfluous, while to repudiate and deny one of these two renders certain existent phenomena unknowable by certain beings.
The second reason for there to be definitely only two valid ways of knowing is that all validly knowable phenomena can be divided into objective and metaphysical entities, in other words nonstatic and static phenomena. A way of knowing must always have and take an object and assume the aspect of something, which then appears in the cognition as a mental hologram. If a valid means of knowing
takes as its appearing object an external objective entity and what appears in the cognition is a reflection of this external entity, it is a bare cognition. If the appearing object of a valid way of knowing is a metaphysical entity, such as a category, and what appears in the cognition is a mental hologram representing this internal entity, it is an inferential cognition. Therefore, as there are specific valid ways for the aspect of each type of entity to arise as what appears in a cognition, it is certain that there need be only two valid ways of knowing.
Chandrakirti, however, asserts four valid ways of knowing in his Clarified Words (Tshig-gsal, Skt. Prasannapada) commentary to the first verse of the first chapter, “Conditions,” of Nagarjuna’s Root Verses for Madhyamaka, Called Discriminating Awareness (dBu-ma rtsa-ba shes-rab, Skt. Prajna-nama-Mulamadhyamakakarika):
(I.1) Nowhere and at no time have functional phenomena been known to arise from self, from other, from both or from no cause. In his explanation of how cognition arises devoid of such impossible ways of arising, Chandrakirti specifies four ways for knowing mundane knowledge that are differentiated according to what they rely on as the condition for them in to arise:
Valid straightforward cognition (mngon-sum tshad-ma, Skt.pratyakshapramana), which arises from an aspect (a mental hologram) of either what Sautrantika calls an objective entity or a metaphysical entity appearing on the consciousness without direct reliance on a line of reasoning
Valid inferential cognition (rjes-dpag tshad-ma, Skt. anumanapramana), with which knowledge of something obscure arises from a line of reasoning that does not deviate from (being applicable to) what is to be established about it
Valid cognition from comprehension of an analogous example (dpe-nyer ‘jal-ba’i tshad-ma, Skt. upamanapramana), which relies on the acquisition of the significance of something not experienced before from something similar to it, such as knowing a gayal wild ox as being similar to an ox – or to use a modern example, knowing the route to your destination from a map
Valid cognition from a scriptural source (lung-gi tshad-ma, agamapramana), which derives from the words of an authoritative source (yid-ches-pa, Skt. apta) who has had straightforward cognition of something beyond observation.
Thus, although Sautrantika would classify this third and fourth way of knowing as types of inferential cognition, Chandrakirti specifies them to illustrate his explanation of how cognition arises from conditions that are neither self, other, both or neither. Chandrakirti’s explanations forms the basis for the unique Prasangika assertions concerning ways of knowing. This fourfold scheme is especially helpful in the context of the Prasangika presentation of the causal conditions for the arising of straightforward cognition of voidness, first conceptual and then non-conceptual. Such cognition may arise from:
Reliance merely on the words of explanation in the scriptural sources.
Further Twofold Divisions of Valid Cognition
Concerning this, the definition of a valid cognition is a fresh, non-fraudulent awareness. When divided, there are two (types): (1) valid bare cognition and (2) inferential cognition. From another (point of view), there are (another) two (types): (1) valid cognition that determination (of what its object is) is self-induced and (2) (valid cognition) that determination (of what its object is) must be induced by another (cognition).
Valid cognition that determination of what its object is is self-induced (rang-las nges-pa) and valid cognition that determination of what its object is must be induced by another cognition (gzhan-las nges-pa) are not separate ways of knowing that are different from valid bare cognition and valid inferential cognition. They are merely a different way of cutting the pie, as it were, of valid knowings into two.
As previously explained, in any cognition, the fallaciousness (mi-slu-ba-nyid) of the cognition and the consciousness and accompanying mental factors of the cognition are established simultaneously as a single substantial entity (grub-sde rdzas-gcig) in the cognition. Thus, they are explicitly apprehended by the bare cognition of the reflexive awareness that accompanies the cognition. The reflexive awareness understands that if the significance (don) of its comprehensible object (gzhal-bya) – in other words, the kind synthesis of what its involved object is that is established on the side of the object – was not ascertained by the cognition of
this object, then valid cognition of the object would not have arisen. In other words, when reflexive awareness cognizes the cognition of an object, it cognizes whether or not the cognition has also ascertained the kind synthesis of what its involved object is that is established as a single substantial entity with the object. If it has ascertained this, then the reflexive awareness cognizes that this determination, this certainty, has been self-induced by the cognition. If the cognition has not ascertained the kind synthesis of its involved object, then the reflexive awareness cognizes that ascertainment of it will need to be made by another cognition.
The non-fraudulence of a cognition, however, can not only be known by the reflexive awareness that accompanies the cognition. It can also be known by the bare cognition or inferential cognition of an object itself, since both the object and the kind synthesis of what it is are established as a single substantial entity on the side of the object and therefore also arise as a single substantial entity on the side of the involved object that appears in the cognition. Through familiarity with repeated cognition of an object, the kind synthesis of what an object is can also be ascertained by the bare cognition or inferential cognition itself.
For example, having repeatedly seen the color red, you can know the non-fraudulence of a visual bare cognition you have of the red color of something as being red. Although the redness of some object is established as being red from the side of the red color, when you look at the object in the dark, your visual bare cognition does not ascertain what the color of the object is. With this visual bare cognition of this red object in the dark, you know that your cognition is not non-fraudulent concerning what color the object is and that you will need to look at it again when the light is on in order to determine its color.
Valid cognition that determination of what its object is is self-induced may occur with either a valid bare cognition or a valid inferential cognition. On the other hand, a valid cognition that determination of what its object is must be induced by another cognition can only occur with valid bare cognition.
You may gain valid knowledge of something by relying on either valid people, valid speech or valid cognitions. An example of a valid person is a Buddha. Valid speech is his teachings, such as those in the first turning of the wheel of Dharma concerning the four noble truths of true suffering, it true origins, its true stopping and true pathway minds leading to its true stopping. Valid cognitions are those of bare cognition and inferential cognition.
Since this division of three valid sources of knowing something is from an etymological point of view alone, this means that they are valid only in the general sense of being correct and non-fraudulent. It does not mean that each fulfills the precise definition of a valid way of knowing. Only valid cognitions satisfy the full definition, whereas valid people and valid speech do not, because they are not awarenesses.