IS PERCEPTION INTENTIONAL?

A PRELIMINARY EXPLORATION OF INTENTIONALITY IN INDIAN PHILOSOPHY

By Georges B. J. Dreyfus

 

    The notion of intentionality is one of the most discussed questions of contemporary Western philosophy of mind, the object of ever-renewed interest and controversies.  Many of these discussions, however, seem to assume the notion of intentionality and then discuss whether this feature constitutes the proper characteristic of the mental or not.  In the context of the study of Indian philosophy, the treatment of intentionality has followed along similar lines, questioning whether this notion applies to the Indian material rather than exploring its possible meanings.  In this essay,1 I will examine some of the questions surrounding intentionality in Indian philosophy by focusing on perception.  In the process, I will explore some of the possible dimensions of intentionality in the Indian philosophical tradition in order to reach a richer understanding of this notion and of the views of the mind in the Indian tradition.  

    In talking about the "mind," it is important to define the term—since it is far from unambiguous.  For most Indian traditions, the mind is not a brain structure or a mechanism for treating information.  It is also not an organ working for a self.  Rather, mind is conceived as a complex cognitive process consisting of a succession of related mental states.  These states are at least in principle phenomenologically available, that is, they can be observed by turning inwardly and observing the way in which we feel, perceive, think, remember, etc.  Indian thinkers describe these mental states as cognizing (jñā) or being aware (buddh) of their object.  Thus, the mental is broadly conceived among traditional Indian thinkers as constituted by a series of mental states that cognize their objects.  

    But this general agreement about the mental quickly breaks down when we turn to a more detailed analysis of the nature and structures of the mind, a topic on which various schools entertain vastly differing views.  Some of these disagreements relate to the ontological status of mental states and the way they relate to other phenomena, particularly physical ones.  Hence, they are similar to well known ideas in the Western tradition, particularly the mind-body dualism that has concerned Western philosophy since Descartes.  But many of the views entertained by Indian thinkers are not easily mapped out in Western terms.  One of my taks in this essay will be to explore the different views of the mind in India, particularly as they revolve around the notion of intentionality.

    In discussing intentionality in the Indian tradition, it may be important to start from the realization that this term is of Western origin and has no direct translation in the Indian context.  Hence, in this discussion we will have to take into account analyses of Indian concepts as well as Western discussions, particularly those taking place in the phenomenological tradition. One of the starting point for such discussion is Brentano's famous statement:

          Every mental phenomenon is characterized by what the scholastics of the Middle

          Ages called the intentional (and also mental) in-existence (Inexistenz) of an object

          (Gegenstand), and what we would call, although not entirely in unambiguous terms,

          the reference to a content, a direction upon an object (by which we are not to

          understand a reality in this case), or an immanent objectivity.2

 

For Brentano, mind is characterized by being directed upon an object.  It bears upon this object, regardless of whether this object exists or not.  We cannot think, wish, or dread unless our mind is directed towards something other than itself.  It is this intentional character that differentiates mental phenomena from other phenomena.  

    Although the gist of Brentano's view is clear, the details of his analysis are less so.  For example, what does Brentano mean when he speaks of "in-existence (Inexistenz) of an object"?  This is one of the questions that has animated many in the phenomenological tradition.  It does seem that Brentano first understood intentionality as a relation internal to consciousness.  The mind is aware directly of an an immanent object through which it contacts the external world.  This view is not, however, without many problems; for it would seem that if that were the case, then we would never be able to experience any object more than once.   We would also not be able to share the same experience with others. Realizing these difficulties, Brentano modified his views and even expressed surprise at being attributed an immanent view of intentionality.  However, it is not clear whether his efforts were entirely successful and hence his followers were left with the task of finding more convincing accounts of intentionality.  

    This was particularly the case for Husserl, whose philosophy is largely devoted to an elucidation of this notion.  For Husserl, the object of intentionality is not internal to consciousness but is the external object itself—whether it exists or not.  Otherwise, we would be absurdly enclosed in a private world of solipsistic representations.  In other words, the intentionality of the mind is not to be accounted for by the presence of some special internal object that consciousness contacts, but rather by a special structure that enables the mind to apprehend the external object. This structure is what Husserl calls the noema, the sense through which the mind apprehends an object even in the case where there is no such object.  The mind is such that it always seems to be as if it were contacting an object, even when there is no such object in reality.  The role of the noema is to account for this "as if."3

    In this essay, I will analyze the views of various Indian schools, particularly Buddhist views of the mind, as they cluster around the notion of intentionality.  I will focus my analysis on Dharmakīrti (an obvious choice in a philosophical analysis of the mind in the context of the Indian Buddhist tradition).  I consider his view of cognition as reflexive and his explanations of perception to be the foundational cognitive mode.  I will show that perception raises some specific problems for Dharmakīrti’s system, which privileges perception and greatly limits its purview. I will also show that his account of the intentionality of perception requires the involvement of language and conceptuality; a solution that it is not without creating systemic tensions, but may provide a very fruitful way of thinking about the ways in which we come to know the external world. In exploring the resources that Dharmakīrti’s system offers for understanding intentionality, I will also argue that a more satisfactory Dharmakīrtian account may need to broaden the notion of intentionality. I will thus examine some of the resources that Buddhist philosophy brings to this task before concluding.

Dharmakīrti's View of the Mind

    In many respects, Dharmakīrti's view of the mind is not different from the standard Buddhist views found in various Abhidharma texts in which the mind is presented as a complex cognitive process consisting of a succession of related momentary mental states.  Each state is said to rise in dependence of preceding moments and give rise to further moments, thus forming a mental stream or continuum (santāna, rgyud), much like James' stream of thought.4 Dharmakīrti also views the mind as being composed of momentary mental states that arise in quick succession.  Each moment of consciousness comes to be and disappears instantaneously, making place for other moments of awareness.  Moreover, each moment apprehends the object that appears to it thereby revealing the object apprehended.  This view is reflected in the definition of consciousness (jñāna) as: "Apprehension of an object is the [defining] characteristic of consciousness."5

    For Dharmakīrti, the mind is characterized by its ability to apprehend objects.  This ability can be understood in several ways, as we will see shortly, but what differentiates the mind from other phenomena is its ability to cognize objects.  Several Abhidharma texts reflect this view and it seems to have been widely accepted within the Indian Buddhist tradition.  For instance, Vasubandhu defines consciousness as "the discrete cognition [of objects]" (vijñānaṃ prativijñaptiḥ).6  This is similar to Asaṅga's definition of consciousness when he says: "What is the characteristic of consciousness?  The characteristic of consciousness is to cognize.  It is through consciousness that one cognizes form, sound, taste, tangible, the mental objects and the various domains."7  Consciousness apprehends its objects and this is what differentiates it from other phenomena.  Thus, it does seem quite clear that in the Indian Buddhist tradition intentionality provides a way to identify the mental. The mind is characterized, for the most part, by its ability to apprehend objects.  

    But this brief assertion does not provide an analysis of the nature of intentionality. What does it mean for a mental state to have an object?  We commonly assume that we have unproblematic access to external objects through our senses and our thoughts.  Investigation shows, however, that this may not be the case.  There are cases of perceptual illusions, as well as states in which we are not deceived, in which the perceptions of individuals vary greatly.  Similarly, we cannot assume unproblematically that our thoughts, even when correct, are in direct contact with reality.  We cannot take for granted the assertion that mental states apprehend their objects, and hence need to examine what enables the mind to cognize the world.  To do this, we turn to the logico-epistemological tradition that was started by Dignāga around 500 CE, and was greatly expanded upon by Dharmakīrti at least a century later.  Its contribution was the explicit formulation of a complete Buddhist logical and epistemological system, in which the primary concern was the nature of knowledge.

    In the Indian context, the issue of knowledge is formulated as the question: what is the nature of valid cognition (pramāṇa), and what are its types? Hindu thinkers tend to present a realist theory, which liberally allows for a diversity of instruments of valid cognition. For example, the Hindu Sāṃkhya school asserts that there are three types of valid sources of knowledge: perception (pratyakṣa), inference (anumāna), and verbal testimony (śabda). The Nyāya, perhaps the most important in the Hindu logico-epistemological tradition, added a fourth type of valid cognition—analogy (upamāna). This fourfold typology was extremely influential and provided the most authoritative epistemological typology in India. Buddhist epistemology, however, rejects these typologies and offers a more restrictive view, limiting knowledge to inference and perception. It is in its examination of inference as a source of knowledge that the Buddhist tradition analyzes reasoning, in particular the conditions necessary for the formation of sound reasons and all their possible types. Hence this tradition is often described, somewhat misleadingly, as “Buddhist logic.”8

    The interpretation of the word pramāṇa is itself a topic of debate among Buddhist and Hindu thinkers. For Hindus, in accordance with its grammatical form, pramāṇa refers to a “means of valid cognition.” This understanding also accords with their view that knowledge is owned by a subject—the self—to whom knowledge is ultimately conveyed. For instance, the Nyāya school asserts that knowledge is a quality of the self—it is only when I become conscious of something that I can be said to actually know it. This view is emphatically rejected by Dharmakīrti, who follows the classical Buddhist line that there is no knowing self, only knowledge. Pramāṇa, therefore, should not be taken in an instrumental sense, but rather as referring to the knowledge-event, the word itself being then interpreted as meaning valid cognition. Dharmakīrti defines valid cognition as:

          . . . That cognition [that is] non-deceptive (avisaṁvādi). Non-deceptiveness [consists]

          in the readiness [for the object] to perform a function.9

 

This statement emphasizes that pramāṇa is not the instrument that a knowing self uses to know things. There is no separate knowing subject, just knowledge— pramāṇa. According to this account, a cognition is valid if, and only if, it is nondeceptive. Dharmakīrti in turn interprets nondeceptiveness as consisting of an object’s readiness to perform a function that relates to the way it is cognized. For example, the nondeceptiveness of a fire is its disposition to burn, and the nondeceptiveness of its perception is its apprehension as burning. This perception is nondeceptive because it practically corresponds to the object’s own causal dispositions, contrary to the apprehension of the fire as cold.

    But the purview of the Buddhist logico-epistemological tradition is not limited to purely epistemological issues and includes discussions of the nature of cognitive events.  In conceptualizing cognition, Dharmakīrti relies on the concept of aspect (ākāra), a notion that goes back to the Sāṃkhya school, but has been accepted by several other schools as well.  The idea behind this position, sākāravāda ("assertion of aspect") in Indian philosophy, is that cognition does not apprehend its object nakedly, but through an aspect—the reflection or imprint left by the object on the mind.  For example, your visual sense consciousness does not directly perceive a blue color, but rather, captures the likeliness of blue as imprinted on cognition.  It is that cognitive form that allows us to differentiate among our experiences.  To be aware of an object does not mean apprehending this object directly, but rather it is having a mental state that has the form of this object and being cognizant of this form.  The aspect is the form or epistemic factor that allows us to distinguish mental episodes.  Without aspects, we would not be able to distinguish a perception of blue, for instance, from a perception of yellow—since we do not perceive yellow directly.  Therefore, the role of the aspect is crucial in Dharmakīrti's system.  Consciousness is not the bare seeing that the direct realist and common sense imagines, but instead the apprehension of an aspect, the cognitive form of the object which stands for this object in the field of consciousness. This aspect, however, is not external to consciousness.  It is the form under which an external object presents itself to consciousness.  And yet it is also the form that consciousness assumes when it perceives its object.  Thus, an aspect is both a representation of an object in consciousness and the consciousness that sees this representation.  

    This double nature of aspects is a consequence of their being intermediaries between the external world and consciousness.  To perform this role, they must partake of both domains.  That is, they must look like external objects and be of the nature of consciousness.  This double nature is implicit in the word åkåra, which means form and quality, that is, the form through which the object is seen.  It also means aspect, in that it is the aspect under which consciousness sees the object.  Åkåra is the aspect of the object in the consciousness as well as the aspected consciousness itself. The implication of this analysis is that perception is inherently reflexive.  Awareness takes on the form of an object and reveals that form by assuming it.  In the process of revealing external things, cognition reveals itself.  

    This view of cognition as self-luminous (svayam prakāśa) and self-presencing, is not unique to Dignāga nor to Dharmakīrti.  It is accepted by other thinkers, particularly Hindu Vedāntins, who identify consciousness as the self and describe it as being "only known to itself" (svayaṃvedya) and self effulgent (svayaṃprabha).10  But for Dignāga and Dharmakīrti, the inherently reflexive character of consciousness is less a consequence of the transcendent and pure nature of the mind (as it is for some of his commentators) than its consisting of the beholding of an internal representation. From one side, consciousness has an externally oriented feature called the objective aspect (grāhyākāra).  This is the form that a mental state assumes under the influence of an external object.  From another side it is the internal knowledge of our own mental states.  Referred to as the subjective aspect (grāhakākāra), this feature ensures that we are aware of the objective aspect, the representation of the object.  However, these two parts do not exist separately.  Rather, each mental state consists of both and is therefore necessarily reflexive.

    The necessary reflexivity of consciousness is understood by Dharmakīrti and his followers as a particular type of perception called self-cognition (svasaṃvedana).  Self-cognition can be compared to what Western philosophers call apperception, namely, the knowledge that we have of our own mental states.  It is important to keep in mind, however, that here apperception does not imply a separate cognition.  For Dharmakīrti, apperception is neither introspective nor reflective—it does not take inner mental states as its objects.  It is the self-cognizing factor of every mental episode which brings us a non-thematic awareness of our mental states.  For Dharmakīrti, reflexivity does not require a separate cognition but is the necessary consequence of his analysis of perception in which a subjective aspect beholds an objective aspect, which represents the external object within the field of consciousness.  Self-cognition is nothing over and above this beholding.  

Dharmakīrti's ideas are not unlike those Western philosophers who have argued that consciousness implies self-consciousness. Such philosophers include (despite their otherwise vast differences) Aristotle, Descartes, Locke, Kant, Husserl, and Sartre. According to Locke, a person is conscious of his own mental states. He defines consciousness as “the perception of what passes in a man’s mind” (Essay Concerning Human Understanding (II: ii, 19). Leibniz, in his New Essays Concerning Human Understanding (II: i, 19), criticizes Locke, pointing out that this view leads to an infinite regress—for if every cognitive act implies self-awareness, self-knowledge must also be accompanied by another awareness, and so on ad infinitum. This regress arises, however, only if knowledge of one’s mental states is assumed to be distinct from knowledge of external objects. This assumption is precisely what Dharmakīrti denies. A consciousness is aware of itself in a nondual way that does not involve the presence of a separate awareness of consciousness. The cognizing person simply knows that she cognizes without the intervention of a separate perception of the cognition. This knowledge is the function of apperception, which thus provides an element of certainty with respect to our mental states. However, apperception does not necessarily validate these states. For instance, one can take oneself to be seeing water without knowing whether that seeing is veridical. In this case, one knows that one has an experience, but one does not know that one knows. The determination of the validity of a cognition is not internal or intrinsic to that cognition, rather, it is to be established by practical investigation.

    Several arguments are presented by Dharmakīrti to establish the reflexive nature of consciousness.11 One of his main arguments concerns the nature of suffering and happiness as it reveals the deeper nature of mental states. For Dharmakīrti, as for Abhidharma literature as well, suffering and happiness are not external to consciousness, but integral to our awareness of external objects. Our perceptions arise with a certain feeling-tone, be it pleasant, unpleasant, or neutral. They are a function of the presence of the mental factor of feeling as described by Abhidharma texts. This feeling needs to be noticed, however, otherwise we would not be aware of how the apprehension of the object feels. Because this noticing cannot be the function of another mental state without incurring the problem of an infinite regress, it must be the mental state apprehending the external object that becomes aware at the same time of the feeling. This conclusion indicates, for Dharmakīrti, the dual nature of mental states.

    In a single mental state, two aspects can be distinguished: the objective aspect (the representation of the external object in consciousness). and the subjective aspect (the apprehension of this appearance, or self-cognition). For Dharmakīrti, a mental state also has two functions. It apprehends an external object (ālambana) and beholds itself. The apprehension of an external object is not direct, but results from the causal influence of the object, which induces cognition to experience (anubhava)—the object’s representation. Therefore, the mind does not experience an external object, but beholds an internal representation that stands for an external object. Cognition cannot be reduced to a process of direct observation, but involves a holding of an inner representation. This beholding is not, however, an apprehension in the usual sense of the word, since the two aspects of a single mental episode are not separate. It is an ‘intimate’ contact, a direct experiencing of the mental state by itself through which we become aware of our mental states at the same time as we perceive things.

    The view of cognition as reflexive does not yield a full account of the mental.  Reflexivity may, however, be a necessary condition for mind.  It explains the direct access that we feel when we have our own mental states, but it not sufficient to account for our cognitions.  In particular, it does not respond to the question: what is it about the mind that enables it to be directed toward external objects?

Is Perception Intentional?

    For Dharmakīrti, the most obvious answer would seem to be found in his theory of perception, which he presents as the foundation of knowledge.  When conjoined with the idea that perception is unmistaken, it would seem that this view, which seems to have much in common with empiricism, should find it easy to account for the way in which perception contacts the world.  The world is given to our senses and is interpreted by our thought process, thus resulting in our knowing the objects of the world. Knowledge, then, can be reduced to a contact of the senses with external reality.  External objects generate internal representations or impressions, which are similar to these external objects and represent them within the field of consciousness.  Knowledge consists of our noticing these representations and interpreting them.  Or, to put this in Dharmakīrtian terms, external objects give rise to aspects that are apprehended by self-cognition. Those cognitive imprints, which are similar to external objects, represent them within the field of consciousness, much like Lockean ideas represent external objects.  

    There are elements of Dharmakīrti's philosophy that correspond to this empiricist view, particularly his sharp dichotomy between conception and perception and his privileging of the latter as the foremost and sole unmistaken valid cognition.  According to Dharmakīrti, a mental state can apprehend its object in two ways. It can apprehend its object positively, engaging in the object as it is. This is the way of perception, which apprehends real objects and is unmistaken).  Or, a mental state can operate by elimination, engaging not the real object as it is but in its conceptual representation.  This is the way of thought, which is not fully and directly grounded in reality but is limited to the realm of fictional constructs and hence mistaken.  Since conception is mistaken, it is only through the intermediary of perception—the only unmistaken access to reality—that conception can be valid under the form of inference (the other allowed type of valid cognition).  This sharp separation between perception and conception is enshrined in Dharmakīrti's definition of perception (Commentary, III: 300.cd) as the cognition that is unmistaken (abhrānta) and free from conceptions (kalpanāpodha). Since perception is unmistaken, and since conception is mistaken, it is perception that must be completely free from conception.

    Given this sharp dichotomy between perception and conception and the privileging of perception, one would expect Dharmakīrti to favor the empiricist analysis of intentionality sketched above. However, Dharmakīrti’s view is far more complex and subtle. Like Sellars, he rejects the idea that reality is directly given, holding that knowledge, even at the perceptual level, does not boil down to an encounter with reality, but requires active categorization. We do not know things by sensing them. Perception does not deliver articulated objects, only impressions. By themselves they are not forms of knowledge, but become so only when they are integrated within our categorical scheme.  For example, when one is hit on the head, that person will initially have an impression and sensation of pain which is not by itself cognitive.  This sensation becomes cognitive only when it is integrated within a conceptual scheme in which it is explained as being an impact on a certain part of our body due to certain objects.  It is only then that the impression of being hit becomes fully intentional or representational.  Prior to this, the impression, or to speak in Dharmakīrti's language, the aspect, does not represent anything.  It only becomes a representation when conceptually interpreted.  

    This conclusion may seem surprising given Dharmakīrti's emphasis on the primacy of perception, but it logically derives from his analysis of the validity of cognitions.  For Dharmakīrti, cognitions are valid if, and only if, they have the ability to lead us towards successful practical actions.  In the case of perception, however, successful practical actions are not as straight forward as one might think. Achieving practical purposes depends on correctly describing the objects we encounter.  It is not enough to see an object which is blue; we must also see it as being blue.  To be a non-deceptive cognition depends largely on the appropriate identification of the object as being this or that.  Perceptions, however, do not identify their objects since they are not conceptual—they cannot categorize their objects, they merely hold them without any determination.  Categorization requires conceptual thought under the form of a judgment.  Such a judgment, which is induced by perception, subsumes its object under an appropriate universal, thereby making it part of the practical world where we deal with long-lasting entities which we conceive as parts of a determined order of things.  

It would seem that the role of perception must be limited to holding passively its object without determining it and inducing appropriate conceptualizations.  For instance, when we sense a blue object, which we categorize as blue, the perceptual aspect (the blue aspect) is not yet a representation, because its apprehension, the perception of blue, is not yet cognitive.  It is only when it is interpreted by a conception that the aspect becomes a full fledged intentional object standing in for an external object.  Now, it would appear as though, Dharmakīrti is bound to maintain that perceptual aspects are transformed retrospectively into representations by conceptions.

    This analysis of perceptual intentionality seems to shift the burden toward conception.  This is obviously problematic, given Dharmakīrti's privileging of perception and the sharp separation between unmistaken perception and mistaken cognition.  Nevertheless, it appears that this shift is necessary within Dharmakīrti's system.  It is not possible to explain intentionality in the full-blown sense of the word without having recourse to conceptuality.  Let me say a few words about this topic before pursuing my analysis of the intentionality of perception.

Thought and Language

    In examining thought (kalpanā), Dharmakīrti postulates a close association with language. In fact, the two can be considered equivalent from an epistemological point of view. Language signifies through conceptual mediation in the same way thought conceives of things. The relation between the two also goes the other way—we do not first understand things independently of linguistic signs and then communicate this understanding to others. Dharmakīrti recognizes a cognitive import to language; through language we identify the particular things we encounter, and in this way we integrate the object into the meaningful world we have constructed. The cognitive import of language is particularly obvious in the acquisition of more complex concepts. In these cases, it is clear that there is nothing in experience that could possibly give rise to these concepts without language. Without linguistic signs, thought cannot keep track of things to any degree of complexity. Dharmakīrti also notes that we usually remember things by recollecting the words associated with those things. Therefore concepts and words mutually depend on one another.

    This close connection between thought and language, inherited from Dignāga, differentiates Dharmakīrti from classical empiricists such as Locke and modern sense-data theorists who believe in what Sellars (1956) describes as the “myth of the given.” Locke, for example, holds that concepts and words are linked through association. The word “tree” acquires its meaning by becoming connected with the idea tree, which is the mental image of a tree. Hence for Locke, the representation of the tree is not formed through language, but is given to sensation (Dharmakīrti’s perception). We understand a tree as a tree through mere acquaintance with its representation without recourse to concepts. Dharmakīrti’s philosophy is quite different, for it emphasizes the constitutive and constructive nature of language. This conception of language is well captured by one of Dharmakīrti’s definitions of thought:

          Conceptual cognition is that consciousness in which representation (literally,

          appearance) is fit to be associated which words.12

 

Thought identifies its object by associating the representation of the object with a word. When we conceive of an object we do not apprehend it directly, but through the mediation of its aspect. Mediation through an aspect is also the case for perception, but here the process of mediation is different. In the case of perception there is a direct causal connection between the object and its representation, but no such link exists for thought. There is no direct causal link between the object and thought, instead there is an extended process of mediation in which linguistic signs figure prominently.

    For Dharmakīrti, the starting point of this process is our encounter with a variety of objects that we experience as being similar or different. We construct concepts in association with linguistic signs in order to capture this sense of experienced similarity and difference. This linguistic association creates a more precise concept in which the representations are made to stand for a commonality that the objects are assumed to possess. For example, we see a variety of trees, and apprehend a similarity between these objects. At this level, our mental representations have yet to yield a concept of tree. The concept of tree is formed when we connect our representations with a socially formed and communicated sign; we assume that they stand for a treeness that we take individual trees to share. In this way experiences gives rise to mental representations, which are transformed into concepts by association with a linguistic sign. The formation of a concept consists of the assumption that mental representations stand for an agreed upon imagined commonality. Thus concepts come to be through the conjunction of the experience of real objects and the social process of language acquisition. Concept formation is connected to reality, albeit in a mediated and highly indirect way. According to this view, concept formation is also considered to be mistaken.

    A concept is based on the association of a mental representation with a term that enables the representation to stand for a property assumed to be shared by various individuals. In Dharmakīrti’s nominalist world of individuals, however, things do not share a common property; the property is rather projected onto them. The property is manufactured when a representation is made to stand for an assumed commonality, which a variety of individuals are mistakenly taken to instantiate. Hence this property is not real; it just a pseudo-entity superimposed (adhyāropa) on individual realities. This property is also not reducible to a general term. In other words, the commonality that we project onto things does not reside in using the same term to designate discrete individuals. Upon analyzing this notion of sameness of terms, we realize that identifying individual terms as being the same presupposes the concept of sameness of meaning, in relation to which the individual terms can be identified. Commonality, then, is not due simply to a term, but requires the formation of concepts on the basis of the mistaken imputation of commonality onto discrete individuals. What does it mean, however, for a concept to be based on an assumed commonality?

    Here Dharmakīrti’s theory must be placed within its proper context—the apoha or exclusion theory of language created by Dignāga. This complex topic is beyond the scope of this essay. Suffice it to say that the apoha theory is a way to explain how language signifies in a world of individuals. Linguistic meaning poses a particularly acute problem for Dignāga and Dharmakīrti, for they are committed to a connotationist view of language, in which the senses have primacy over reference. However, such a view is difficult to hold in a nominalist ontology that disallows abstract entities such as meaning.13

    The apoha theory tries to solve this conundrum by arguing that language does not describe reality positively through universals, but negatively by exclusion. Language is primarily meaningful, but this does not mean that there are real senses. Instead, we posit agreed upon fictions that we construct for the sake of categorizing the world according to our purposes. Therefore “cow” does not describe Bessie through the mediation of a real universal (cowness), but by excluding a particular (Bessie) from the class of noncow. Matilal describes Dignāga’s view this way:

          Each name, as Dignāga understands, dichotomizes the universe into two: those to

          which the name can be applied and those to which it cannot be applied. The function

          of a name is to exclude the object from the class of those objects to which it cannot

          be applied. One might say that the function of a name is to locate the object outside

          of the class of those to which it cannot be applied.14

 

Although linguistic form suggests that we subsume an individual under a property, analysis reveals that words merely exclude objects from being included in a class to which they do not belong. The function of a name is to locate negatively an object within a conceptual sphere. The impression that words positively capture the nature of objects is misleading.

    The apoha theory was immediately attacked by Hindu thinkers such as Kumārila and Uddyotakara, who raised strong objections. One of which was that it is counterintuitive—we do not perceive ourselves to eliminate noncows when we conceive of cows. Dharmakīrti’s theory of concept formation is in many ways an attempt to answer these attacks. It argues that the apoha theory is not psychological, but epistemological. In conceiving of objects we do not directly eliminate other objects, but instead rely on a representation that is made to stand in for an assumed commonality shared by several individuals. It is this fictional commonality that is the result of an exclusion. There is nothing over and above particular individuals, which are categorized on the basis of their being excluded from what they are not. The concept that has been formed in an essentially negative way is projected onto real things. In the process of making judgments such as “this is a tree,” the real differences that exist between the different trees come to be ignored and the similarities are reified into a common universal property—which is nothing but a socially agreed upon fiction.

    The eliminative nature of thought and language is psychologically revealed when we examine the learning process. The word “cow,” for instance, is not learned only through a definition, but by a process of elimination. We can give a definition of “cow,” but that definition works only if its elements are already known. For instance, we can define cows as “animals having dewlaps, horns, and so on” (the traditional definition in Indian philosophy). But how do we know what counts as a dewlap? It is not just by pointing to the neck of a cow, but by eliminating the cases that do not fit. In this way, we establish a dichotomy between those animals that fit, and those animals and things that do not. And it is on this basis of a negative dichotomy that we construct a fictive property of cowness. This construction is not groundless, however, but proceeds through an indirect causal connection with reality. Concepts are not formed a priori, but elaborated as a result of experiences.

    Dharmakīrti’s solution to the problem of thought and meaning is thus to argue that in a world bereft of real abstract entities (properties), there are only constructed intensional (linguistic) pseudo-entities, and this construction is based on experience—perception. This grounding in perception ensures that although conception is mistaken, in the way reviewed above, it is neither baseless nor random, and hence can lead to the formation of concepts that will be attuned to the causal capacities of particulars.

 

The Intentionality of Perception Revisited

    We are now in a better situation to appreciate the subtleties, as well as the difficulties of Dharmakīrti's account.  Because he is a consistent nominalist, Dharmakīrti holds that that in a world bereft of real abstract entities there are only constructed intensional pseudo-entities.  He also holds that this construction is based on perception.  It is this grounding in perception that ensures that although conception is mistaken, it is neither baseless nor random.  For Dharmakīrti perception is not by itself fully cognitive and intentional. Perception does not determine the situation as it is cognitively understood, but brings about certain forms of conceptual activity in which we apply or withdraw certain concepts we have previously learned.  These forms of memory are necessarily conceptual and hence negative and mistaken.  For example, the judgment "this blue pot is beautiful" does not come about through mere acquaintance with the object, but requires a conceptual elaboration in which concepts are formed by excluding contrary assumptions such as "this is not blue," "this is not a pot," and "this is not beautiful."  This conceptual activity is not arbitrary, it arises within the limitations imposed by experience and yet does not reflect reality directly.  Our assertions and negations, which constitute our knowledge, are based indirectly on the reality we perceive; the truth of our conceptions is based on their being connected with perception.

    This view is not unlike that of some of the interpreters of Husserl's view of the role of noema.  According to his view, knowledge cannot be reduced to contact with the senses, which provide us with data but not with knowledge.  To have knowledge, we need to bring these data together by a structure, the noema, that makes them into the data of an object.  Otherwise we would be forever limited to the solipsism of the present moment.  There are, however, disagreements among Husserl scholars concerning the question of the status of the noema.  The noema is neither the external object as it is given nor the representation of the external object in consciousness, like Locke might assert.  Rather, the noema is something like the content of consciousness, the sense through which consciousness intends its object.  But if this is so, then how can the noema be part of the perceptual act?  It is here that interpreters seem to part ways along lines that can be roughly grouped into two camps.  There are the so-called West Coast interpretations of Dreyfus and Follesdal, who, relying on Husserl's earlier writings (particularly—but not only—the Logical Investigations), argue that the noema is an abstract and non-sensible entity that can be apprehended only conceptually.  This view is rejected by the so-called East Coast interpretations of Sokolovsky, Cobb-Stevens and Zahavi, who argue that it does not represent Husserl's mature view.  For them, the noema is not conceptual but apprehended by the senses, though not given.15  This essay is not the place to decide who is right in this matter, at best an arduous task completely outside of my competences. Nevertheless, I think that this difference raises a question that is significant for understanding intentionality in the Buddhist epistemological tradition.

    Dharmkīrti's view is similar in certain respects to the so-called West Coast interpretations.  For both, the senses provide us with the data (the aspect), but those do not by themselves provide us with knowledge.  For this we need to group data together and conceive of them as pertaining to a single object. It is only through the intervention of thought that the mind can be said to be intentional.  Perception gives the datum as it is, however, it is not able to determine what it is. Conception determines and understands the object by subsuming it under a universal, but does not see it.  Knowledge of the external world necessitates both seeing and conceiving and, therefore, requires the cooperation of these two cognitive elements which are powerless in isolation.  

    Nevertheless, perception is not to be considered completely blank or purely passive. It has the intentional function of delivering impressions that we take in and organize through our conceptual schemes. Perception can be said to have what I would call a phenomenal intentionality. This kind of intentionality may be revealed in certain forms of meditative experiences. Dharmakīrti alludes to such experiences when he describes a form of meditation in which we empty our mind without closing it completely to the external world.16 In this state of liminal awareness, things appear to us but we do not identify them—we merely let them be. When we come out of this stage, the usual flow of concepts returns, and with it the conceptualization that allows us to identify things as being this or that. This experience shows, Dharmakīrti argues, that identification is not perceptual, but the result of conceptualization. Moreover, in such a meditative state, perception takes place but not conceptualization. Hence, perception is a nonconceptual sensing onto which interpretations are added. Due to the speed of the mental process, the untrained person cannot differentiate conceptual from nonconceptual cognitions. It is only on special occasions, such as in some form of meditation, that a clear differentiation can be made. Therefore, when the flow of thought gradually subsides, we are able to reach a state in which there is a bare sensing of things. From within this state, what we call shapes and colors are seen as they are delivered to our senses without the adjunctions of conceptual interpretations. When one gradually emerges from such a nonconceptual state, the flow of thoughts gradually reappears, and we are able to make judgments about what we saw during our meditation. One is then also able to make a clear differentiation between the products of thoughts and the bare delivery of the senses, and distinguish cognitive from phenomenal intentionality.

Conclusion: Is This Enough?

    Dharmakīrti's analysis of the cognitive process provides a rich account of the elements whose cooperation is required in order to explain intentionality. One of the great merits of this analysis is to disentangle the processes through which we come to know the world—explaining the role of perception as a way to contact the world while emphasizing the role of conceptual categorization in the formation of practical knowledge. This account, nevertheless, suffers from several problems—most of which derive from a commitments to an epistemological dualism and a sparse ontology.  But there are other problems as well. For instance, the question of whether Dharmakīrti's analysis of intentionality is rich enough to account for all the cognitive processes known to the Buddhist tradition of his time?  

    To answer this question, we need to go back to the Abhidharma, the matrix of Buddhist philosophy in India.  As we already saw, Dharmakīrti's view mostly agrees with the Abhidharma analysis of the mind.  But is his view able to accomodate all the significant insights about the mind contained in the Abhidharma?  In most versions of the Abhidharma, there are six types of consciousness: five born from the five physical senses (sight, hearing, smell, taste and touch) and mental cognition. Each type of sensory cognition is produced in dependence on a sensory basis (one of the five physical senses, and an object).  This awareness arises momentarily and ceases immediately to be replaced by another moment of awareness, etc. . .  The sixth type of awareness is mental.  It is considered by the Abhidharma as a sense, like the five physical senses, though there are disagreements about its basis.17

    Some Abhidharmists, such as Asaṅga, argue that these six types of consciousness do not exhaust all the possible forms of awareness.  To this list, they add two types of awareness: the store-consciousness (ālaya-vijñāna, kun gzhi rnam shes), which is of interest to us here, and afflictive mentation (kliṣṭa-manas, nyon yid), which does not need to detain us here.18  The idea of a store-consciousness is based on a distinction between the six types of awareness, which are described as manifest cognitive awareness (pravṛtti-vijñāna, 'jug shes), and a more continuous and less manifest form of awareness, the store-consciousness.  This consciousness is subliminal and hence we usually do not notice it.  It is only in special circumstances such as fainting that its presence can be noticed or at least inferred.  This consciousness holds all basic habits, tendencies, propensities and karmic latencies accumulated by the individual, thus providing a greater degree of continuity.  

    This type of consciousness is not readily amenable to Dharmakīrti's analysis since it is a cognition and presumably therefore intentional.  But it is hard to understand how such intentionality can be explained within the confine of Dharmakīrti's epistemology.  The store-consciousness does not seem to be cognitive, either directly like conception, or indirectly like perception.  Because it does not capture any feature of its objects, it cannot be said to determine an object like conception.  Since it is subliminal, it does not have any phenomenal content, and hence it is unable induce any categorization.  How then, can it be intentional?  

    To respond to this question in detail would necessitate an analysis that goes well beyond the purview of this essay, but it seems quite clear that at least two avenues could be pursued here. One solution would be to argue that such a basic cognition is not intentional and hence that intentionality is not a criterion of the mental.  The difficult question would then be: what does make such a cognition mental?  Here again, various avenues may be available to answer this question. If we were to use the conceptual resources offered by Dharmakīrti and his tradition, we could argue that this cognition is mental because it is reflexive, thereby making reflexivity rather than intentionality being the criterion of the mental.  The difficulty would then be to explain the reflexivity of a cognition that is usually not directly accessible to us (except perhaps in certain meditative states).  

    Another avenue would be to extend the concept of intentionality.  The store-consciousness is not intentional in the sense of having an object it is about, either directly, or even indirectly.  Rather its intentionality consists of its having a dispositional ability to generate more explicit cognitive states.  Some Western phenomenologists distinguish between the object directed intentionality and operative or dispositional intentionality.19  Whereas the former is what we usually mean by intentionality, the latter is a non-reflective tacit sensibility, a kind of purely passive cognitive level that makes us ready to respond cognitively to the world though it is not by itself explicitly cognitive.  For example, when we walk we have a kind of subliminal sense of our body and the territory on which we are stepping.  And yet, we do not take this place as an explicit object that we place in a broader conceptual framework.  Rather, we have a vague sense of the place where we stand—an inchoate perceptual field that acquires a particular configuration only through being solicited by some particular perceptual clues.  In our example, we suddenly feel our step losing its stability and focus on the ground on which we are standing.  This explicit intentional focus does not come out of nothing but emerges out of a more basic form of intentionality, an openness to the world which though being inchoate is nevertheless structured by implicit abilities, preferences, and expectations.  Out of such a dispositional orientation, a more focused and explicit awareness emerges, a kind of uncovering and focusing of the perceptual field.  

    It seems that this kind of operational intentionality provides an interesting avenue to explore the cognitive nature of the store-consciousness.  Although Dharmakīrti himself never broaches this topic, it is an important subject without which no Buddhist account of the cognitive process can claim to be complete. It is true that the idea of the store-consciousness is connected to a particular tradition, the Yogacāra.  Since Dharmakīrti does seem to belong to this tradition, it could be argued that he is not committed to each and every of its tenets.  But the idea of a long lasting, basic subliminal form of awareness, a kind of degree zero of consciousness, goes well beyond the limits of the Yogacāra school and its particular view of the store-consciousness.  Most of the schools of Indian Buddhism are committed to the existence of such a form of more continuous consciousness, which performs important functions within the Buddhist conceptual universe.  Different traditions describe this continuous consciousness in different ways, but there is overwhelming, though not universal, agreement among Indian thinkers that the consistency of Buddhist philosophy requires the presence of such a form of consciousness.  Moreover, it does seem that a full account of the mental will require an explanation of some of the subliminal mental states that are of great importance in the meditative practices of several Buddhist traditions.  Hence, it seems it would be hard for a Dharmakīrtian account of the mind to avoid completely this question.

 

Georges Dreyfus

Williams College

Winter 2006