歷來，致力於會通康德哲學與佛家哲學的學者們（主要包含Stcherbatsky、K.C. Bhattacharyya、牟宗三，或許還可包含一些日本僧侶背景的哲學家–宇井伯壽、和辻哲郎），都不約而同地站在佛家的立場，指出康德不應該拒絕「智的直覺」，認為在某種特殊的狀態下，「智的直覺」可以直接體證「物自身」，而且認為不只認識對於對象表象之外的本體具有效力，「物自身」、乃至於「自由」，都不如同康德所宣稱，在認知上「不可以有任何 positive account」，而應該可被正面證成。
我主張知識論不應該帶有任何形上學與本體論的預設，而該讓知識論單純作為知識論，姑且稱之為「批判知識論」。更進一步限定，批判知識論主張：一、認識不是一種實際的動作；二、認識依循的因果關係並非經驗的因果關係，而是另外一種形式的（只具形式作用）、自由的（沒有前因、自主肇因）的因果關係，稱之為「自由的因果」；三、認識的因果並不因為不依循經驗的因果關係，而在效力上不能為「積極的 positive」，只是，積極效力不必然意謂著任何特殊的存有狀態。依循著這個觀念，我支持兩個傳統中的「非主流觀點」，並且願意在文本上（康德「Third Antinomy」，陳那「自證理論」）以及哲學思辨上提供辯駁。就此觀點，不只許多因為本體論詮釋而衍生的歧異可以獲得調解，而且知識論與二者的實踐計畫也將因此觀點而獲得更明確的「正位」。
簡言之，「認識能力」、「自由」與「物自身」作為一種預設，並非一種本體的預設、並不依循著自然的因果律，因此它們並不會成為一種認識的對象與結果，對於經驗世界沒有實質地影響，因此這種預設並不會對空性原則造成困難，它們並不需要超越的實在性。但是，若缺乏這種預設，而經驗世界在哲學考量上將無法成立。不同於 Allison 將這種預設看作是一種「單純理論性的」、「系統需求而被迫虛設」的「理性要求」，我主張這個預設的形式效力與實際因果關係的存有效力是同等的（同樣積極的 positive），只是兩者性質上完全「不同」：認識成為事實之時，就是認識的條件被完滿之時，而兩種因果關係同等地都是該認識條件的要求項目。接受經驗實在性，就必須同時接受先驗觀念性–誠如康德一向的主張，同時，我們也應該如此來理解佛家的二諦論，以及龍樹的名句「以有空義故一切法得成」。
統整一下，我所主張的路線，既不認為康德需要接受「智的直覺」，也不以為佛家真的可以主張「智的直覺」（唯識不必然與中觀有究竟的教義分歧），然而，我們也不因此而認同「物自身等不可以有任何 positive account」。換言之，在兩個傳統中，「追求一種超越的認識，可以去認識對象本然的本體」這樣的目的並不適當，物自身、認識的因果關係、認識主體三者，都不需要是種超越的、本體的抑或是實際的存有。「智的直覺」將造成經驗世界在哲學考量上的困難（中觀對唯識的非難多半是朝著這個部份而來），但是「自由」與「物自身」等作為一種先驗觀念，既不是一種先驗幻想（與認識條件相悖的單純概念堆疊的產物），也不是一種消極的預設。先驗觀念，是絕對的肇因，既無前因，也無後果，但是是認識結果的必要前在條件之一，無需要任何形上的或本體的基礎，但認識條件的完滿需要先驗觀念的參與。那麼，「本體論」，就只能是一種「作為認識結果」的本體論，不會是「因」或者「條件」上的研究工作，而「本體」只是一種經驗與理念的雜揉的產物，既沒有超越的狀態，也因為感性條件的限制，不會以「整體」的型態呈現在我們的經驗範圍之中，但是其整體性與必然性，只因為經驗與理性的必然雜揉而在「經驗範圍」中「完全合理」，但也僅止於「合理」，具有「可證性」但不具有「實證性」。
Transcendental Logic and Spiritual Development – Following Dignāga’s and Kant’s Critical Epistemology (PhD. Diss. Proposal, 2014 May): PDF download
Over the past years, the scholars who tried to converge Kant’s philosophy and Buddhist philosophy (including Stcherbatsky, K. C. Bhattacharyya, MOU Zongshan and perhaps some Japanese monk-philosophers UI Hakuju and WATSUJI Tetsurō) coincidently criticized Kant from the perspective of Buddhist thinkers that Kant should not reject intellectual intuition. They, with their oriental resources, suggested that in some extraordinary situation, intellectual intuition can directly know things in themselves and that the cognitive capacity can remain effective even beyond the scope of appearance while the idea of the thing in itself as much as the idea of freedom is indeed more than what Kant had claimed to be without any positive account in cognition.
Such comment, no matter in terms of Kant or Buddhism, actually confused epistemology with ontology in the background. In other words, the commentators understood epistemology with metaphysical and ontological assumption, and then understood both Kantian epistemology and Buddhist epistemology with the confusion. The confused understanding, however, has actually been the mainstream one in both sides, but it is not the case that this understanding has never been questioned in each commentary history. Also, in both fields, the urge exclaiming let epistemology be simple epistemology can be observed in their recent development (Henry Allison in Kant; Dan Arnold, Yao Zhihua, Chu Junjie in Buddhism). In my research, I have indicated materials in the original texts and the classical commentaries supporting the coming-up challenge. This has to be especially for the Buddhist side, because, on the other side, Kant is obviously in favor of the non-mainstream understanding. Kant directly rejected the epistemology with the metaphysical and ontological assumption of “transcendental realism,” and the mainstream commentators actually did not agree, or had problems with such a position of Kant. Besides, I think pure philosophical consideration would suggest any epistemology with ontological assumption cannot hold, either.
Further on, the Buddhist thinkers held various opinions on the efficacy of the intellect (prajñā) on the thing in itself. This has led to the predicament that the role of epistemology in understanding and practicing Buddha’s teachings remains indeterminable for a long time. Those who held on the principle of emptiness (Mādhyamaka) thought the transcendent efficacy would undermine the position that rejects any transcendent reality. The others (Yogācāra) thought epistemology was more appropriate to unfold the “meaning” of emptiness but failed to propose any satisfying answer to the challenge about the contradiction between the cognitive efficacy and the principle of emptiness. I think the reason of the discrepancy between the two sides and the weak response of Yogācāra rests at the impotent notice and development of the “critical nature” of Buddhist epistemology (“critical” means to suspend any ontological assumption). The ontological interpretation of Dignāga’s epistemology, especially his theory of self-awareness (svasaṃvitti), in Candrakīrti’s Mādhyamaka influential response and Dharmakīrti’s inside influential response, I think, has to be responsible for the lack of this development.
I hold epistemology should be critical without any metaphysical and ontological assumption, and epistemology should remain simple epistemology. I call it “critical epistemology.” To specify, critical epistemology holds following three points. (1) Cognition is not real action. (2) Cognition does not follow empirical causality but is in another kind of causal relation which is formal (yielding forms, not matters) and free (spontaneous and autonomous). I call it “causality of freedom” comparing to “causality of nature.” (3) It is not because cognition does not follow natural causal laws that we can hold that its efficacy cannot be “positive”; nonetheless, its positivity does not necessarily imply any particular status of existence. With this idea, I will argue for the “non-mainstream” interpretation in both traditions, textually (“third antinomy” in Kant and “theory of self-awareness” in Dignāga) and philosophically. Hopefully, with the idea, the discrepancies caused by the ontological interpretation could be reconciled and the role of epistemology in the practical projects of both sides could be finally determined.
In short, cognitive capacities, freedom and the thing in itself as presumptions are not ontological presumptions – they are simply about cognition itself (the condition of cognition). They do not follow natural causal laws. They cannot be the objects and the results of cognition and they do not really affect the empirical world. Then they would not cause any difficulty for the principle of emptiness. However, they are needed. Without the presumptions, the philosophical consideration to the empirical world cannot hold. Besides, unlike Allison’s opinion to understand these presumptions as simply theoretical demands of reason forced by system, I think the formal efficacy is as much effective (positive) as real efficacy. But they are different in kind. Both as required items in the condition of cognition, when cognition is realized, i.e., when the condition of cognition is satisfied, they both have to be effective. If we have to accept empirical reality, we have to accept transcendental ideality, too – as Kant always claimed. Also, I suggest we should try to understand Nāgārjuna’s “if you make sense of emptiness, everything makes sense 以有空義故一切法得成.”
Let me be clear, I hold Kant needn’t accept intellectual intuition, and Buddhism does not really hold intellectual intuition, either (i.e., there should be no original discrepancy between Yogācāra and Mādhyamaka). However, it is not followed that the thing in itself, etc. do not bear any positive account in cognition. In other words, pursuing some kind of transcendent cognition of the transcendent objects in themselves is not appropriate at all in both traditions, and the thing in itself, the causal relation of cognition and the subject of cognition needn’t be any transcendent, ontological or real existences. Intellectual intuition is rejected because it causes problems in the philosophical consideration of the empirical world (this is exactly the target of Mādhyamaka’s attacks on Yogācāra). However, freedom and the thing in itself, as transcendental ideas, are not transcendental illusions (simple conceptual constructions which do not fit the condition of cognition). They are not negative, theoretical presumptions, either. Transcendental ideas are the absolute cause. Without any pre-condition and without any real result, they are in the condition which makes the result of cognition possible by holding the parts of the condition in a unity. There need not any metaphysical or ontological basis for them, but the satisfaction of the condition of cognition needs transcendental ideas. Following the above, ontology can best be the ontology merely about the results of cognition, not about the cause or the condition of cognition. “Ontology” (or the Chinese concept Ben-ti 本體) is after all just the product of the mixture of experience and ideas. There’s no transcendent status of it, and because of the limit of our sensibility, there’s no real status of it in totality. But because of the necessary “participation” of reason in experience, the idea(ontology)’s totality and necessity in experience is totally reasonable – and at best only reasonable: the idea can be proved in the study of the condition of experience but not evidenced with any sensational proofs.