Non-Innate A Priori Knowledge in Avicenna

I. The Background Story

In his "The Empiricism of Avicenna," Dimitri Gutas interprets Avicenna as an empiricist.¹ He analyzes Avicennian 'principles of syllogism' and claims that none of them are a priori. Moreover, regarding awwalīyāt and fiṭrīyāt—which are two groups of such principles—Gutas suggests that "[i]t appears that both kinds of propositions would be analytic, in Kantian terms. As for Locke, they would be what he called 'trifling.'"² In my first comment in this issue, I disagreed with this view and argued that these two groups of propositions are a priori in the Kantian sense. Assenting to their truth is internal to the intellect and independent of empirical information. I also argued that at least some fiṭrīyāt are synthetic, rather than analytic. So Avicenna's epistemology accommodates instances of non-analytic knowledge that are independent of sense experience. This casts serious doubt on the plausibility of describing Avicenna as a full-blown empiricist. Gutas has responded to my note and claimed that the discussion of a priority, in its Kantian sense, is irrelevant to Avicenna's epistemology. He also defends the claim that awwalīyāt and fiṭrīyāt are both trifling and experience-dependent.

Before offering my rejoinder, I would like to thank Professor Gutas for his willingness to respond to my argument and for helping to clarify what is at stake. In the following section, I present a version of my argument, which, compared to the version developed in my previous comment, has less emphasis on Kantian notions. I do this to show that the strength of my argument does not depend on the application of Kantian language. In section three, I discuss Gutas' objections and explain why they fail. Section four concludes.

II. Avicennian Sense-Experience-Independent Propositions

For Avicenna, acquiring a piece of knowledge is either forming a concept (taṣawwur) or assenting to the truth of a proposition (taṣdīq). A concept is dependent on sense experience if and only if no human can grasp it without having specific sense experiences (i.e., without having specific experiences of the external world). Similarly, a proposition is dependent on sense experience if and only if even after grasping its conceptual components we still need to have specific sense experiences to assent to its truth. But according to Avicenna there are propositions whose truth can be assented to independent of all sense experiences, even if their conceptual components are all dependent on sense experience. Avicenna believes that such propositions are imposed by the intellect on its own. This is exactly what Avicenna says about awwalīyāt and fiṭrīyāt. They could be constituted from concepts that are dependent on sense experience. Nevertheless, after grasping these concepts we do not need to receive further information through our senses to assent to the truth of these propositions. So, in this sense awwalīyāt and fiṭrīyāt are sense-experience-independent propositions that are constituted from sense-experience-dependent concepts. This idea can easily be extracted from some of the passages that Gutas himself has quoted:

An example of [awwalīyāt] is [the proposition] that "the whole is greater than the part." One does not come into possession of this either from sense perception (ḥiss) or induction (istiqrā') or anything else. To be sure, it is possible that sense perception provides him with a concept (taṣawwur) of "the whole," "the greater," and "the part," but acknowledging the truth (taṣdīq) of this proposition comes from his nature (jibilla)…³

It might be the case that no one can grasp the concepts 'the whole', 'the greater', and 'the part' without having any sense experience. If so, these concepts are sense-experience-dependent. However, after grasping them we can assent to the truth of the awwalī proposition 'the whole is greater than the part' without appealing to any further sense experience. So, the assent to the truth of this proposition is not dependent on sense experience.⁴

The same story applies to a fiṭrī proposition. Its conceptual components might be grasped through some specific sense experiences. But, after grasping them, the truth of the proposition can be acknowledged by the natural operation of the intellect and without appealing to further sense experiences. Consider, for example, the proposition 'four is even'. It is constituted from sense-experience-dependent concepts. No one can obtain the concepts 'four' or 'even' independent of sense experiences.⁵ However, after grasping these concepts, as soon as one considers them, the middle term—i.e., 'divisible by two'—will appear in the intellect, and the truth of the proposition will instantaneously be acknowledged. Put otherwise, if one considers the concepts 'four' and 'even', then the intellect will immediately provide a syllogism for establishing the fiṭrī proposition 'four is even', and, accordingly, the truth of this proposition will automatically be acknowledged. That is why such propositions are called 'propositions with built-in syllogisms' (qad.āyā qiyāsātuhā ma'ahā or muqaddamāt fiṭrīyāt al-qiyās).⁶

Since awwalīyāt and fiṭrīyāt are composed of sense-experience-dependent concepts, they are not innate or inborn. The truth of such propositions is not already acknowledged upon birth. They should come about in the intellect as the child grows up. This is because it is obviously impossible to have sense-experience-dependent concepts upon birth (i.e., before having any sense experience). Moreover, the truth of a proposition cannot be assented to until its conceptual components are grasped. This indicates that awwalīyāt and fiṭrīyāt are not innate. In brief, awwalīyāt and fiṭrīyāt are noninnate propositions that are composed of sense-experience-dependent concepts, but their truth can be assented to without appealing to any sense experience.

Moreover, in contrast with Gutas' proposal, Avicennian fiṭrīyāt are neither analytic in Kantian terms nor trifling in Lockean terms. In this sense, their status is different from awwalīyāt that are analytic or, equivalently, trifling. For Kant a proposition is analytic if and only if its predicate is contained in its subject. To put it in a more Aristotelian-Avicennian form, a proposition is analytic if and only if its major term is contained in its minor term. But by contrast with awwalīyāt, fiṭrīyāt have middle terms and in principle need to be demonstrated by a syllogism. It does not matter that such a syllogism will instantaneously be made by the intellect. The mere need for such a (hidden) syllogism indicates that the major term of a fiṭrī proposition is not contained in its minor term. Thus, it cannot be an analytic proposition. Although awwalīyāt are analytic, fiṭrīyāt are synthetic. Furthermore, the Kantian distinction between synthetic and analytic propositions is construed by many scholars as an (at least) approximate counterpart to the Lockean distinction between instructive and trifling propositions.⁷ If so, fiṭrīyāt would be instructive, rather than trifling as Gutas suggests.

Finally, what Kant means by a priority and a posteriority can be captured by the following theses:

1. 1. A piece of knowledge is a priori if and only if it is independent of sense experience.

2. 2. A piece of knowledge is a posteriori if and only if it is not a priori.

Relying on these definitions, awwalīyāt and fiṭrīyāt are non-innate propositions that are composed of a posteriori concepts. Nevertheless, they are still a priori propositions in the sense that assenting to their truth is independent of sense experience.

III. Gutas' Objections

I do not find it necessary to address everything about which Gutas and I disagree. I restrict my discussion only to two of the most important objections that, if successful, can disarm my argument. The objections are articulated in Gutas' words, though the phrases in brackets are mine.

Objection 1. [The proposition 'four is even'] is essentially a tautology: in the minor, "four" is the species of the genus "divisible in two equal parts" and thus included in it (a "verbal" proposition as Locke says in my note 41); and in the major, "divisible in two equal parts" is the definition of "even," and thus completely identical with it, tautologically.

Reply 1. According to Avicenna, neither 'divisible in two equal parts' nor 'even' is a genus of number. Nor is 'even' essential/per se (dātī) to 'four.' Rather, it is only an essential/per se accident ('araḍdātī) of 'four'. If 'even' was essential to 'four', it would also have been essential to all even numbers greater than four. So, everyone who fully grasps, for example, the concept 'thirty-six' would know that it is even. But Avicenna believes that in principle it is possible to grasp the concept of such a number without knowing that it is even. Moreover, it is impossible that 'even' is essential to only some but not all even numbers. Therefore, it would be essential neither to a number like thirty-six, nor to four.⁸ But it is difficult to see how it is possible that a concept P is contained in a concept S without the former being essential to the latter. As a result, it seems to be untenable to say that the concept 'even' is contained in the concept 'four'. But if the predicate of a proposition is not contained in its subject, it cannot be either analytic in the Kantian sense or trifling (or verbal) in the Lockean sense. This observation is perfectly compatible with the examples Locke gives for his trifling propositions. For instance, he believes that 'Lead is a metal' is trifling because 'Lead' is a species of the genus 'metal'. So, 'metal' is contained in 'Lead' because the latter is essential to the former. Thus, as soon as one fully grasps these concepts one acknowledges the truth of the proposition 'Lead is metal'.⁹ However, as we saw, the proposition 'four is even' in principle needs to be demonstrated. It does not matter how quickly its demonstration can be made by the intellect. The mere fact that it requires to be demonstrated shows that it is not trifling.

Objection 2. Zarepour writes: "Coming-about-in-mind is not the same notion as non-a priority." [But] we have no way to understand what Zarepour is talking about other than to assume that "non-a priority" means "coming from experience," in which case Zarepour's statement would be "Coming-about-inmind is not the same notion as coming from experience." In this case it is clear that Zarepour disagrees with Avicenna and disapproves of his use of "coming about," ḥaṣala, because Avicenna says precisely this (L8 in my article): "As for the primary notions (al-awā'il) that come about for him/occur to him (taḥṣulu lahū), they come to pass from induction (istiqrā'), which is testing and proving (tajriba), and from Experience (šahāda)."

Reply 2. If we understand 'experience' as referring to anything that happens to a human after birth, then, of course, coming-about-in-mind is the same notion as coming from experience. In this sense, every instance of noninnate knowledge (i.e., what we grasp after birth) is by definition experience-dependent. In other words, according to this definition of 'experience', non-innate knowledge would mean nothing but empirical knowledge. On the other hand, if we understand 'experience' as referring to what happens to us due to contact with the extramental (or outside) world, then it is not obvious that coming-about-in-mind is the same as coming from experience. According to the latter understanding, the term 'experience' refers only to our sense experiences (i.e., the experiences through which we grasp something from the external world). Hence, if 'experience' is understood in this way, then it would not be a trivial fact that every instance of non-innate knowledge is empirical (i.e., dependent on sense experience). Indeed, my argument shows that according to Avicenna awwalīyāt and fiṭrīyāt are noninnate propositions that are independent of sense experience.

The flying man argument provides perhaps even more obvious evidence that, for Avicenna, even non-innate knowledge can be independent of sense experience. This thought experiment is intended to show that we can obtain knowledge of our existence without having any sense experience. It seems to be incontrovertible to see the flying man argument as evidence that, according to Avicenna, the concepts 'I' (or 'self') and 'existence' and the proposition 'I exist' are all independent of sense experience, even if they are not innate. The flying man argument is compatible with the claim that there is no instance of innate knowledge, but contradicts the claim that all instances of knowledge are empirical (in the sense of being senseexperience-dependent). To sum up, there are two different senses of 'experience'. If it refers to anything that happens to us after birth, then every non-innate piece of knowledge would, by definition, be empirical. But if what 'experience' refers to is restricted to sense experiences, then Avicenna does not believe that every non-innate piece of knowledge is necessarily empirical.

Gutas says that Avicenna disagrees with the claim that coming-about-inmind is not the same notion as coming from experience. This interpretation is justifiable only if 'experience' is understood in its broad sense, which includes anything that happens to us after birth. Gutas takes 'Experience' ("with a capital E, as a perceptive faculty") as a translation for what Avicenna calls 'mušāhada.' As Gutas reports, Avicenna uses the latter term to refer to "our direct apprehension of the outside world and internal self."¹⁰ If so, mušāhada includes every perceptive phenomenon that happens to us after birth. Thus, it is true that for Avicenna every non-innate piece of knowledge is experience/mušāhada-based. Coupling this with the fact that according to Avicenna there is no instance of innate knowledge, we can conclude with the following thesis:

(III) For Avicenna every instance of knowledge is experience/mušāhadabased.

I agree with Gutas that (III) is defensible. However, my argument shows that the following thesis is false:

(IV) For Avicenna every instance of knowledge is sense-experience-based (i.e., dependent on our experiences of the outside world).

Now the problem is this: Gutas seems to think that (III) shows that Avicenna is an empiricist in the modern sense. I disagree. Attribution of empiricism to Avicenna is defensible only if we can establish (IV). But I showed that (IV) is false; therefore, we cannot call Avicenna an empiricist in the modern sense. Indeed, what Gutas has shown is, at most, that Avicenna does not endorse innatism. But non-innatism is not equivalent to empiricism. Innatism implies non-empiricism, but the other way around does not hold. For sure, a philosopher is not an empiricist if s/he believes that we have cognitive tools through which we can obtain non-innate pieces of knowledge that are independent of sense experience, and Avicenna is such a philosopher.

One might object that Gutas is not really concerned with the modern sense of empiricism. To justify this claim one could even refer to Gutas' caveat that "in the study of the history of philosophy one cannot explain one philosopher through another because the terms of discourse and referents are never exactly the same in the two philosophers." So it might be concluded that what Gutas refers to by 'empiricism' is irrelevant to the modern sense of the term. However, there seems to be a not easily resolvable tension between this conclusion and Gutas' claim that "[I]t is obvious that Avicenna had a systematic epistemology, which he articulated a number of times, and empiricism—in the Western European sense as it was later to appear in the work of John Locke—played a commanding role in it."¹¹

IV. Conclusion

It seems that we are dealing with a dilemma. If we understand empiricism as equivalent to non-innatism, then we are justified in describing Avicenna as an empiricist. However, this sense of empiricism seems to be irrelevant to its modern sense. On the other hand, if, in accordance with the modern discourse, we understand empiricism as the thesis that all instances of knowledge hinge on our experiences of the outside world, then Avicenna cannot be described as an empiricist. This is because for him there are instances of knowledge that are independent of sense experience. So, on either of the horns of the dilemma, it is implausible to describe Avicenna as an empiricist in the modern sense.