Inferential cognition is the comprehension of an obscure fact through reliance on a correct line of reasoning as its basis. When divided, three types are explained: inferential cognitions based on (1) the force of the actuality of phenomena, (2) what is well-known and (3) conviction.
 To know explicitly something obscure and not readily obvious, you must rely on the valid support of either the force of the actuality of phenomena (dngos-po’i stobs), what is well-known (grags-pa) or conviction (yid-ches). For instance, when your neighbor is making a great deal of noise you may become annoyed and impatient because it is not obvious that sound is impermanent. However, if you rely on the force of the actuality of phenomena, you can deduce and prove to yourself that this noise will pass simply because it is something produced– in other words, something that has arisen immediately from effort. To do so you must rely on the three factors of applicability to the topic, pervasion and negative pervasion. This noise was something produced; if something is produced it is pervasive that it is impermanent and will pass, and it is pervasive that if something is permanent it is not something that was produced. Therefore, through inferential cognition based on the force of the actuality of phenomena (dngos-stobs rjes-dpag), you can be certain that this noise will also pass. With such valid knowledge, you can then control your anger.
 Just as Westerners have traditionally seen a “man in the moon” when looking at its craters, Indians have seen a “rabbit in the moon.” When in Sanskrit and Tibetan literature you read about “that which has a rabbit,” these words do not refer to their obvious, literal meaning. You know that such a literary allusion refers to the moon through an inferential cognition based on what is well-known (grags-pa’i rjes-dpag) or a popular convention. In Western literature, you know that a man’s best friend is his dog through a similar valid means. This is also the method by which you know what any word means when you hear the sound of someone uttering it, for all words are popular conventions.
 There are certain things that are extremely obscure and only when you become a Buddha can you have bare cognition of them. Before that, you must rely on your conviction in the Buddhas’ scriptural texts to know them at all. Since Buddhas are valid persons and what they have said is valid speech, you can infer that by relying on them you will have valid cognition. Thus, through an inferential cognition based on conviction (yid-ches rjes-dpag), you can be sure that prosperity is the result of previously practiced generosity.
Inferential cognition and valid inferential cognition are to be known as mutually inclusive.
Therefore, all inferential cognitions relying on correct lines of reasoning are valid.
Valid Cognition in Which Determination Is Self-Induced or Must Be Induced by Another Cognition
Valid cognition in which determination (of its object is) is self-induced or must be induced by another (cognition) is the valid cognition that decisive determination can be induced through its own power or that of another (cognition concerning the fact) that if the significance of its own comprehensible object were not established on top of (it, cognition of this object as having that significance) could not arise.
Valid cognition that determination of its object is self-induced (rang-las nges-kyi tshad-ma) and valid cognition that determination (of its object) must be induced by another cognition (gzhan-las nges-kyi tshad-ma) amount to validly knowing that it either is or is not self-evident what something is. For instance, there is a fire in the distance. When you look at it, you can either cognize it as simply being a red object or as being a fire. When you see it in the former way, but not in the latter, you can validly know two things about this object.
- You can validly know that your decisive determination that this object is red has derived from this bare visual cognition itself, which means to realize that it is self-evident that the object is red. You validly know that if its significance as red were not established on it, you could not have seen it as red.
- Furthermore, if from where you are standing you cannot be sure whether this object is a fire or a red cloth, you can also validly cognize that to decisively determine whether it is a fire, you will have to rely on the power of another cognition. In other words, you realize that only by having a closer look can you become certain of the fact that if it were not a fire you could not see it as a fire. Thus, when seeing this item as merely a red object, you can validly know that it is not self-evident that it is a fire.
Another example is seeing a tree at a distance. That it is a tree is self-evident and you validly know that your cognition of it as such has been self-induced. But it is not self-evident that it is an oak and you are aware that a decisive determination of whether it is an oak or an elm will have to be induced by another cognition. Only when you come closer will you know for sure
Valid cognition that determination (of its object) is self-induced can be any of five (types of) valid cognition: two (kinds) of valid cognition – (1) bare cognition by reflexive awareness or (2) yogic bare cognition – (3) inferential cognition, or (valid sensory bare cognition) of (4) something with its functioning (also) appearing or (5) something with which one is familiar.
 When your faculty of reflexive awareness has fresh bare cognition of a state of consciousness, it also cognizes that its decisiveness about its object is self-induced. No further cognition is necessary.  The same is true of valid yogic bare cognition cognizing either subtle impermanence or the coarse or subtle selflessness of persons. If the decisiveness of what was apprehended by such cognition were not self-induced, you could not have apprehended it at all.
 With valid inferential cognition you reach a correct conclusion from a valid line of reasoning. Nothing further is required to know this conclusion, therefore your valid cognition is self-induced.
 When you have sensory bare cognition of the manifestation of something’s ability to produce an effect, such as a fire’s consumption of fuel, you are directly perceiving what is happening. If it required another cognition to know what it was you were perceiving, then you could not say you were actually witnessing the manifestation of such an effect. You would not know specifically what you were perceiving at all.
 If you have seen your friend’s son every day and are totally familiar with him, then whenever you have valid sensory bare cognition of him, even at a distance, it is self-evident that he is the son of your friend. If you are a master repairman, then whenever you see a broken appliance you know immediately what is wrong and how to repair it. Because of your complete familiarity, your decisive determination of what the problem is is self-induced without the need of further cognition.
When divided from the point of view of etymology, valid cognition that determination (of its object) must be induced by another (cognition) is of three (types): bare cognition of something (1) for the first time, (2) when one’s mind is unheedful and (3) having a cause for deception.
 When you see an utpala lotus for the first time, it is self-evident that it is a blue flower, but not what specific kind it is. To realize when you see this that you will need further information and cognition to identify it is an example of this first type of knowing that valid cognition of its species will have to be induced by another cognition. It will have to be “other-induced.”
 The second type occurs, for instance, when someone says something to you while you are engrossed in thinking about something else. Aware that you have heard something, you realize that it will have to be repeated for you to validly know what has been said. Such valid cognition often occurs with non-determining cognition.
 When you see a mirage of water in a desert and realize that you will need to have a closer look to be certain of what you have seen, this is an example of knowing, when your present cognition is affected by a cause for deception, that your decisive determination of what it is will have to be induced by another cognition
These last two types of cognition are valid in the sense that with them you realize that what you are perceiving is not self-evident. But because the cognitions themselves are non-determining or distorted, they are valid only in an etymological sense and not in an actual one.
There are also the valid cognitions that determination of (1) what the appearance (of its object is) is self-induced, but what in truth it is will have to be induced by another cognition, (2) what (its object) is in general is self-induced, but what it is specifically will have to be induced by another cognition, and (3) whether something has even appeared will have to be induced by another cognition. Although such (cognitions) have been explained, care is needed in differentiating which are actually (valid) and which are (only) nominally so.
 An example of this first type is seeing something red in the distance. What appears to your sensory bare cognition, a red color, is self-evidently red, but that this is in fact a fire is not self-evident. Thus, when seeing this object, you can validly know both that your determination of how it appears – it appears as a red object – has been self-induced, but what it is in truth will have to be induced by another cognition.
 The second is seeing a tree in the distance. What it is in general, a tree, is self-evident. To know specifically that it is an oak, you will have to go closer.
 You see a man on a hill out of the corner of your eye. Unsure if you actually have seen a man, you realize you will have to look at the hill more carefully to be certain. This is an example of the third type. Another one is seeing someone and, wondering if you have ever seen the person before, realizing that you will need to have another look to be sure.
These first two are actual valid ways of knowing something. But to realize that the determination of whether something has even appeared to you will have to be induced by another cognition is only nominally called valid. In actuality it is non-determining or may even be distorted.
Although it is pervasive that valid cognition that determination (of its object) must be induced by another (cognition) is a valid cognition, yet because that object, which one validly cognized that determination (of what it is) must be induced by another cognition, may (itself) not be validly cognized, precise detail is required concerning the pervasions and so forth.
Thus, you may invalidly know something, such as a mirage, but validly realize you will have to look at it again to be certain what it is. This is valid from the point of view of correctly knowing that the determination will have to be induced by another cognition. But because it is based on a distorted cognition, this cannot actually be considered valid.