In response to a question in a recent interview concerning his use of analytic philosophy in the book Oneself as Another, Ricoeur says that a passage through the outside is necessary, given the intimist tendency of phenomenology. It is a passage justified by the fundamental fact that the body is both my body and a body among bodies; therefore, the approach of objectification is not to be ignored.1 If the analytic tradition, which Ricoeur calls "the thought from outside" (la pensee du dehors) is so important, one cannot help but ask: why his comments in the book on analytic philosophy are far more negative than positive? Why his point is always to show its inadequacy with regard to the understanding of the self? On the other hand, is the double reading of the corporeal phenomenon not already suggested by Gabriel Marcel whom Ricoeur always regards as his master? Does Gabriel Marcel not see body in terms of mystery-that in which the distinction between the interior and the exterior loses its meaning? Could a deeper understanding of Oneself as Another be attained by a detour through Ricoeur's interpretation of Marcel's distinction between mystery and problem?
This essay is written precisely with the purpose of examining Ricoeur's main works on Marcel, in order to demonstrate that there is really a significant connection between Marcel's thought and Oneself as Another, so that a Marcelian reading of Ricoeur through the latter's own account of the former may be established. To accomplish this goal, I shall consider in depth the critical remarks made by Ricoeur in "Gabriel Marcel and Phenomenology" in the 70s and then relate them to his argument in "
primaire et 
seconde chez Gabriel Marcel" (Primary Reflection and Secondary Reflection in Gabriel Marcel) written later in the 80s before Oneself as Another was published. Ricoeur's earliest work Gabriel Marcel et Karl Kaspers (Gabriel Marcel and Karl Jaspers) will also be consulted on certain points.2 Both of the aforementioned articles are concluded with the suggestion that the relation between mystery and problem has to be understood in terms of the dialectic between secondary reflection and primary reflection. My contention is that this suggestion is in fact the principle of Ricoeur's methodology in Oneself as Another; it allows us to see more about the necessity of the analytic philosophy in the book. Along the way of my exposition, I also want to show that there are other important features in Oneself as Another which can be traced back to Marcel's philosophy.
From the Characterizable to the Uncharacterizable
The most important comments of Ricoeur on Marcel's philosophy are found in the article entitled "Gabriel Marcel and Phenomenology", originally presented in a colloquium on Marcel's philosophy in 1973 just a few months before Marcel's death. In this article, Ricoeur compares Marcel's philosophical method with Husserl's. The focus of his exposition is on Marcel's attitude towards conceptualization. He starts with the point that both Marcel and Husserl maintain the value of conceptualization. In both philosophies, there is an inherent tendency of refusing any "system", but at the same time there is a persistent concern for subtle distinctions and clarity of thought. In order to make comparison with Husserl, Ricoeur brings up Marcel's "Outlines of a Phenomenology of Having" in Being and Having where he distinguishes "what one has" from "what one is". Ricoeur remarks that by trying to make conceptual distinctions between these two phenomena, Marcel shows the non-psychological character of his approach which is not unlike that of Husserl's. In Marcel's attempt to clarify the notion of having-as-possession (l'avoir-possession), a genuinely eidetic style of analysis can be observed in which meaning is directly read from well-chosen examples. Ricoeur also indicates that in his distinction between the qui and the quid, Marcel does not fail to take advantage of the significant relations suggested by ordinary language (GMP 473/55).
All the above features in Marcel's phenomenology of having enable us to put him on the same track as Husserl's until at a point in the text where Marcel speaks about "reduction".3 Marcel states that the point of his analysis is not a reduction, rather it has to do with the presence of an opaque and irreducible datum which resists our full possession. This little word "reduction", marks the profound difference Marcel and Husserl, according to Ricoeur. He admits that in the context where the word appears, "reduction" does not convey the same meaning as in Husserl. The irreducible of Marcel is the primordial dimension of being which "eludes the framework of an idea that one can have and therefore can circumscribe and dominate intellectually" (GMP 473/55), whereas in Husserl reduction is a notion connected with the epoche and designates the subject's withdrawal of his or her natural attitude to the world. Yet such an idea of the irreducible, which arises from the opposition of being and having, has the effect of moving the whole analysis "away from the plane of notional distinctions-from the eidetic plane, in Husserl's terms-to a more existential plane" (GMP 473/55).
For Marcel, "having" denotes a global way of being which is made possible by one's own body. My body, as the mediation between myself and the world, creates a tension between interiority and exteriority, manifested by the link between the desire to have and the fear to lose. But what I claim to possess and attach to exercises a tyranny over me; it devours me eventually. "Having as such seems to have a tendency to destroy and lose itself in the very thing it began by possessing, but which now absorbs the master who thought he controlled it" (Marcel, Being and Having, 164). Among my possessions are my own body, things, ideas and opinions-even characterization is a kind of possession. The findings of the phenomenology of having drive it to a reflection on the conditions of characterization in general: what makes characterization possible? An object can be characterized only when it is placed at a certain distance in front of an uninvolved observer. Hence, characterization relies on a pretension of being able to cut oneself off from the living links with things and stand before them as mere observer and dominator. Nevertheless, this is precisely the condition under which eidetic description operates. Therefore, Ricoeur remarks that "the idea that being is uncharacterizable brings an end to eidetic phenomenology, which cannot help appearing to be prompted by the will to characterize" (GMP 474/56).
This divergence from eidetic phenomenology continues in the famous distinction between mystery and problem. Treating something as a problem is to see it as data placed before me as if I were not implied in it. Mystery is rather that in which the distinction between in me and before me loses its meaning. Hence, we have the having, the characterizable and the problematizable on the one hand, and being, the uncharacterizable and mystery on the other; but the relation between them is again not characterizable. In Marcel's phenomenology, there is an ascending dialectic from the examination of examples to the recognition of the irreducible-irreducible to characterization (GMP 475/57). Ricoeur mentions that although Marcel's phenomenology stays close to Husserl's in the use of description, eidetic analysis and even imaginative variations, but this inward transcending movement from the characterizable towards the uncharacterizable seems to be lacking in Husserl's phenomenology. Marcel's phenomenology of having starts with examples and descriptions but the internal dialectic of the inquiry eventually turns back upon the conditions under which the whole inquiry begins. This ascending dialectic is for Ricoeur the most essential feature of the Marcelian style of thought which he illustrates with a quotation from "On the Ontological Mystery": "A mystery is a problem which encroaches upon its own data, invading them, as it were, and thereby transcending itself as a simple problem".4 This quotation is so dear to Ricoeur that it appears in almost all of his works on Marcel (GMP 475/57, GMKJ 361, L2 66, 95). And it is also in the interpretation of this statement that I discover the major disagreement between Marcel and Ricoeur; it shows both Ricoeur's dependence on Marcel and his critique of the master. This point will be further elaborated as my exposition proceeds. An initial comparison of the two possible interpretations of the quotation can be made by a detour through the way how the master and the disciple understand the term "transcendence". A more Marcelian interpretation may be established on the meaning of the word "transcendence" which Marcel himself defines in Being and Having as "
rather than Aufhebung" (Marcel, Being and Having, 119). In that case, the statement quoted above would be the last one we can make before ending up with the ineffable. On the other hand, referring to the same passage in an early work, namely Gabriel Marcel et Karl Jaspers, Ricoeur remarks that if transcendence is seen as "rather than Aufhebung", then metaphysics would be a discipline that "does not aspire to maintain a tension, but to solve it" (GMKJ 269). Later we will see that the difference between Marcel and Ricoeur lies precisely in whether a certain tension should be acknowledged between the lower and the higher levels of transcendence, and whether the relation between problem and mystery is just a simple either-or.
Ricoeur explains the difference between the two philosophers in terms of their "initial gestures" by which they enter into philosophy. For Husserl, the initial gesture is reduction whereby the subject suspends its belief of the natural world. The benefit of reduction is twofold. First, the objectivity of the object is revealed as the identical meaning towards which different intentional aims of consciousness converge. Secondly, the subjectivity of the subject is revealed as an intentional consciousness that is caught up in a temporal flux in which it is capable of retaining and anticipating its own identity in the flux. This twofold benefit of reduction characterizes the phenomenology of the early Husserl as "a kind of reflection, a descriptive analysis, applied to the correlations established between the structures of the object and those of the subject" (GMP 476/58). The object, or noema, is that which intended by the mind; the subject, or noesis, is that which signifies by intending meaning. Ricoeur remarks that Husserl's initial gesture of reduction, which favours a clear distinction between the subject and the object, is diametrically opposed to Marcel's. The latter starts with "situation" which he defines in the beginning of The Mystery of Being as "something in which I find myself involved".5 For Ricoeur, to be involved excludes first and fundamentally "both the distance characteristic of reduction and the promotion of a 'disinterested spectator', the very subject of phenomenology" (GMP 476/58). The situation does not only affect the subject from outside but also qualifies it from within. The opposition of the outer and the inner loses its meaning, and along with it, the typically Husserlian correlation of the noematic and the noetic is called into question. One may notice that in Ricoeur's Oneself as Another, the process of the determination of the self also begins with a situation, namely, the situation of interlocution. The self first appears as the one of whom the interlocutors speak (OA 31/44)