Phenomenology, abduction, and argument: avoiding an ostrich epistemology

Phenomenology has been described as a “non-argumentocentric” way of doing philosophy, reflecting that the philosophical focus is on generating adequate descriptions of experience. But it should not be described as an argument-free zone, regardless of whether this is intended as a descriptive claim about the work of the “usual suspects” or a normative claim about how phenomenology ought to be properly practiced. If phenomenology is always at least partly in the business of arguments, then it is worth giving further attention to the role and form of phenomenological argumentation, how it interacts with its more strictly descriptive component, and the status of phenomenological claims regarding conditions for various kinds of experience. I contend that different versions of phenomenological reasoning encroach upon argument forms that are commonly thought to be antithetical to phenomenology, notably abductive reasoning, understood in terms of its role in both hypothesis generation and in terms of justification. This paper identifies two main steps to making this case. The first step takes seriously the consequences of the intrinsically dialectical aspect of phenomenology in intersection with other modes of philosophy, the natural attitude, and non-philosophy. The second step focuses on transcendental reflection and arguments about the conditions/structures they contain. Together, these two steps aim to rescue phenomenology from the objection that it has an “ostrich epistemology” with regard to the ostensible purity of description, the intuition of essences, or the “conditions” ascertained through transcendental reflection.

Phenomenology has been characterized as a "non-argumentocentric" way of doing philosophy (Glendining, 2007;cf. Crowell, 1999), reflecting that the philosophical focus is on generating adequate descriptions of lived experience and thereby returning to the "things themselves". It is also sometimes criticised by philosophical opponents for this same feature. But it would be a mistake to conclude that phenomenology is an "argument free" zone, whether that is intended as a descriptive claim about the work of the "usual suspects", or a normative claim regarding how phenomenology ought to be properly practiced (i.e., strictly as description, not argument, nor explanation). If phenomenology is always at least partly in the business of arguments, then it is worth giving further attention to the form of "phenomenological argumentation". It is also worth attending to how it interacts with its more descriptive component, in which the phenomenological philosopher intuits or directly "sees" that this or that is 'necessary', 'essential', and/or 'constitutive' in regard to a particular experience. I contend that when we do this, we are led to recognise that some of the major forms of phenomenological reasoning encroach upon argument forms that are commonly thought to be antithetical to it, notably abductive reasoning in both the Peircean and contemporary senses of an inference to the best explanation (that is, in terms of hypothesis generation and in terms of justification). I will define abduction in more detail below. For now, it is sufficient to note that abductive reasoning is neither strictly deductive (and thus does not give strict necessity), nor a kind of enumerative induction that involves statistical generalisation. The same can be said about phenomenology too, as we will see, but my emphasis on the role of abduction within phenomenology nonetheless challenges the contemporary orthodoxy which strictly differentiates transcendental phenomenology from both abduction and any sort of inference to the best explanation (Gallagher & Zahavi, 2008b, p. 90;cf. Crowell, 2013, p. 28).
This paper will proceed as follows. Section 1 introduces the idea of phenomenology via certain influential and programmatic statements about the method. I also outline the dialectical motivation for undertaking phenomenology, which suggests that phenomenological description always has a theoretical and cultural background and motivation, with tacit argumentative premises. While it might be maintained that these motivations (and implicit argument structures that are the conditions of possibility of phenomenology) are extraneous to the phenomenological investigations proper, it is difficult to operationalise this claim when we examine the work of the "usual suspects": the texts are themselves a hybrid mix. Even if we move to consider the genesis of any phenomenological reflections and insights themselves -ostensibly prior to any transcription in language 1 -some related issues arise, including the basic decision to attempt to undertake phenomenological description, and then in the decision as to when that task is sufficiently complete (i.e., essential and fulfilled). Prima facie, any move of "methodological separation" encounters the problem of what has been called "ostrich epistemology" by Alvin Goldman and subsequently many others. This phrase stands for an epistemology in which we have knowledge, but hide from it, as if, like ostriches, we put our heads in the ground when faced with an epistemic threat, thereby avoiding conflicting information or evidence 2 . In more phenomenological language, this is the problem of the phenomenologist needing to wear multiple "hats": that is, with and without the reduction that "brackets" existing or prior knowledge, to enable non-biased descriptions of experience and its conditions (Overgaard, 2015). I argue there is no easy solution here, other than Merleau-Ponty's 3 , which is that the reduction is a heuristic and imperfectly achieved, as he states in his Preface to Phenomenology of Perception (2012), and that phenomenology is always dialectical. I add to his view the claim that it is an unavoidable abductive element that stymies any strict methodological separatist interpretation.
Section 2 focuses on the role of transcendental arguments in phenomenology. These arguments postulate necessary conditions or presuppositions in a way that involves argumentation and can be considered in their logical form. Transcendental arguments have been standardly taken to be deductive in form. These arguments are based on a Kantian model and are also shaped by the formal analytic reception of them (see Chase & Reynolds 2010). Regarding phenomenology, however, the 'transcendental' (and the associated argument structure) is not focused on the conditions of any possible experience in the same way. Rather, it is more like a "material a priori" as Husserl puts it: that is, conditions for actual experience of creatures like us (Taylor, 1975). There is even said to be an experience of the transcendental, and hence non-deductive access, which runs counter to Kant's conception. The appropriate formal representation of these arguments is difficult to categorise, and thus eminently debateable, but I argue that it appears more like an inference to the best explanation (hereafter IBE) or other abductive arguments than a deductive one, and thus does not provide necessity (for all times and places and all forms of conscious experience). While this conclusion may run counter to some key methodological ideas and the express declarations of some phenomenologists, my argument here is also "reconstructive". There are difficulties that continue to beset transcendental arguments on the deductive conception, especially if they are thought to provide necessary and sufficient conditions and ampliative new knowledge. I argue in what follows that if we want to interpret phenomenology charitably and avoid these issues, we are forced to admit a surprising link between modest transcendental reasoning and abduction, including IBE.

Phenomenological description and eidetic intuition
Phenomenological description is central to any work of philosophy that warrants the name phenomenology. There is no version of phenomenology sans description. It is necessary, though, to understand the specifics of phenomenological description and how it works. We can begin by noting that an emphasis on description is enshrined 2 Although commonly associated with Goldman, the remark seems to have been made by him at a conference and I cannot find a published work of his where it occurs. As an aside, and a point of fact, ostriches do not actually hide from danger in this fashion -rather, they dig in the ground with their beaks to secure their eggs. 3 As Merleau-Ponty observes, "it would be a very romantic way of showing one's love for reason to base its reign on the disavowal of acquired knowledge" (Merleau-Ponty, 1964, p. 24). in Husserl's methodological "principle of all principles" for phenomenology: "everything originarily … offered to us in 'intuition' is to be accepted simply as what it is presented as being, but also only within the limits in which it is presented there" (Husserl,Ideas,Sect. 24). This methodological point holds that careful descriptions of that which is "given" in intuition in different kinds of experience is an essential component of phenomenology. In addition, other sorts of putative evidence derive their potential justification from their ability to be apprehended in such a self-present or fulfilled intuition. The phenomenologist aims to describe those experiences and intuitive insights as faithfully as possible. I will consider the nature of such intuitions as I proceed, both in terms of how we arrive at them (which I will suggest is framed by abductive hunches) and the moment of insight itself. I will begin by noting that we are not natural-born phenomenologists. This is true even if there is a sense in which embracing the phenomenological method can be understood as a process of relearning in which we "waken to wonder", as Fink famously put it and Merleau-Ponty glosses in his Preface (2012). There is a paradox at the heart of the effort to undertake phenomenological description: it is a capacity that we have, perhaps even our "ownmost" possibility, and yet it has been concealed or "forgotten" due to the prejudice of objectivism and our captivation by what Husserl and others call the "natural attitude". Indeed, contrary to the natural attitude in which we assume existence, and various common sense or doxic elements embedded within the lifeworld, there is even a sense in which phenomenology is expressly "unnatural" and requires at least some training and enculturation. To put it mildly, it is difficult to refrain from theorising when describing. It is difficult to refrain from existence commitments and certain sorts of explanation (i.e., causal) or basic forms of abductive reasoning, including some of those associated with common sense and the natural attitude.
There are several open questions about how to describe phenomenological experience without betraying it. It might be a matter of ridding ourselves of certain limitations of scope (the natural attitude) and then describing what we "see" directly as essential or constitutive. It might be understood as categorial or eidetic intuition akin to revelation, or more akin to a reflection. It might be a form of reasoning that is also argumentative and comes with prior doctrines concerning method, evidence, and epistemology. Do we simply see essential structures, even effortfully? Or are essential structures disclosed (or posited) through reflection in ways that might continue to warrant the name 'argument' or 'explanation' (Erklärung), rather than being "given" in accordance with the "principle of all principles"? These questions are difficult to resolve, and there is no consensus about their resolution amongst the "family" of selfdescribed phenomenologists. The consensus is perhaps only that concrete description of lived-experience is important to philosophy and human inquiry (Carman, 2006, 99).
Related issues arise in Heidegger's reflections on phenomenology. Despite his own hermeneutic turn, Heidegger still famously begins Being and time by stipulating that the aim of phenomenology is to "let that which shows itself be seen from itself in the very way in which it shows itself from itself" (Heidegger 1994). Part of this "showing" involves an activity of description, perhaps even almost tautologically so 4 . While the differences between Heidegger and Husserl are much debated (cf. Crowell, 2012;Overgaard 2004), there is some continuity in these methodological remarks about phenomenology. And, without simply endorsing this, there is a prima facie sense in which these remarks about phenomenology can be (and have been) taken to support a view of phenomenology as invested in an "ostrich epistemology".
After all, this identifiably phenomenological method of description is often directly opposed to other methods and ways of knowing. Phenomenology thereby derives its identity at least partly from what it is (allegedly) not. A common negative refrain about phenomenology is that it is not mere introspection, nor qualitative psychology; instead, it is radically different in kind. Nor are phenomenology's descriptive methods able to be reduced to any other major philosophical method. In his Preface to Phenomenology of perception (2012), Merleau-Ponty puts forward the orthodox phenomenological view that describing is not analysing, nor explaining. There is some merit to this neither/nor move, of course, but it is perhaps better understood as a difference in kind rather than one of degree, at least if the arguments of this paper hold. We will return to this, but we are entitled to ask for a positive characterisation of what this method of description is exactly. Some argue that this is ultimately left unclarified in Merleau-Ponty's Preface (cf. Smyth, 2014), if not the Phenomenology as a whole. And even if this were not so, questions remain. Notably, it remains to be determined how the ostensible phenomenological moment (or phenomenological acts, which are then described) intersects with the many parts of the text in which Merleau-Ponty both explains and analyses. These are major parts of Merleau-Ponty's book, including parts where he engages the explanatory inadequacies of alternative views, notably gestalt psychology, and the infamous opposition of empiricism and intellectualism.
Stepping back from Merleau-Ponty, we might generalise to claim that phenomenology is always motivated by, and occurs in, a particular historical and cultural context. This means that any particular use of phenomenology starts from somewhere, in media res in a hermeneutic context that involves various sorts of reasoning, both implicit and explicit. These presuppositions are conditions of the possibility of phenomenology, and may be heuristically "bracketed" in order to facilitate a better understanding of the things themselves, but they are never wholly bracketed without remainder, I would argue. This is because the context of hypothesis generation and that of discovery (phenomenological or not) are not wholly separable, with abductive reasoning involved in both. I will argue for this as I proceed. But one way to begin to motivate this claim is through the simple recognition that, in addition to commencing from a motivated position, any strictly phenomenological description must also end, both as a text and as a series of discrete phenomenological acts or inquiries. Even on a suitably nuanced understanding of phenomenological description, it is not clear when any description will have been sufficiently exhaustive and when it will have reached its end point, whether by Husserl's technique of imaginative variation or not. It is not clear how we know that we have exhausted the modal possibilities 5 , or captured the essential ones. Any such decision seems to invoke tacit criteria and something akin to an untutored hypothesis that remains operative in guiding the description (this is the Peircean abduction). More phenomenologically, we could call this the "forestructure of our understanding", as Heidegger does. With Merleau-Ponty, we might say that [e]mpiricism cannot see that we need to know what we are looking for, otherwise we would not be looking for it, and intellectualism fails to see that we need to be ignorant of what we are looking for, or equally again we should not be searching. (Merleau-Ponty, 2012, p. 28).
It seems to me that a certain phenomenological understanding of pure description, along with a strict methodological separation of phenomenological description from the messy mix of our socio-cultural situation and life-world, would be a form of empiricism, even if less atomistic than the usual sort.
It is also the case that not all descriptions are equally philosophically illuminative. It is not that any "hetero-phenomenological catalogue" of first-person experiences suffices to count as phenomenological. Any judgment or reflection that enables us to distinguish those proffered in the natural attitude from those more deeply embedded in the phenomenological attitude, presupposes a comparison between the two. The criteria for this are thin and difficult to intersubjectively establish, but for the phenomenologist him or herself there is such a comparative judgment. It also evokes the tribunal of experience, aiming to approximate to it, and then to generate intersubjective agreement, even if this has perhaps not transpired in quite the way that Husserl anticipated. I will return to that issue of the history of phenomenological heresies, as Paul Ricoeur has put it, but for now my point is just this: if this descriptive effort of phenomenology transforms doxastic experience and sheds new light on it in some ampliative way, then a genuine question arises about whether phenomenological description still warrants the name of description simpliciter. Judgments about this will differ, but it is clear that much is built into the descriptive effort envisioned in the name of phenomenology, perhaps enough to begin to compromise the strict opposition with other argument forms.
This oppositional methodological stance helps to set phenomenology in contradistinction to other ways of doing philosophy. To put this point slightly differently, this opposition helps to identify phenomenology with philosophy, in contradistinction to an array of non-philosophical discourses and modes of inquiry. While this gesture was socio-politically useful to guard against naturalistic encroachment from psychologism, it risks over-reach, perhaps even what Dominique Janicaud has described as the "over-bidding" of phenomenology (Janicaud 2000, 89-107). Indeed, if we may deploy but a single true method, then it might also seem that we have no need of other methods, including analysis and explanation. On this view, phenomenology appears to verge on advocating for a naïve return to innocence and an ostrich epistemology. There are also ways of interpreting such a practice as leading to methodological solipsism, as Dennett (1991) does, accusing phenomenology of "papal infallibility". In the end, I think we can resist these and related charges, but doing so takes us into the territory of abduction and denies any neat "methodological separatism".
It is worth noting that the "pure description" and "non-argumentocentric" ideas sit uneasily with an examination of what goes on in phenomenological writings. The major works associated with the phenomenological tradition appear to represent a mix of methodologies. For all of those producing this work, there also remains a question about the dialectical motivation for doing phenomenology. This involves argumentative premises, even if they are tacit. What is being described, and why is it being described? Why seek a philosophy that "returns to the things themselves"? There are inevitably reasons and motivations for this -philosophical "hunches" even, search strategies, and explanatory conjectures. These may be of the following sort. To grasp this or that phenomena (i.e., sympathy) a phenomenological approach is envisaged as well-suited, perhaps better than other approaches. This again seems tantamount to Peircean abduction. Sometimes these hunches might be more theoretically informed, thus becoming more like an inference to the best explanation. A phenomenological approach to sympathy might be considered likely to be fruitful in comparison to alternative attempts to understand sympathy. These attempts might: (i) be ensconced in what Husserl calls the "natural attitude"; (ii) rely on scientific explanation of a reductive sort; (iii) depend solely on linguistic or conceptual analysis; or (iv) start from a prior, elaborate metaphysics and then look to the world of experience for confirming instances.
Here, it is helpful to reconsider the problem of the phenomenologist needing to wear multiple "hats" -with and without the reduction (cf. Overgaard, 2015). Phenomenologists bracket (but they know that they bracket) and they know something about what they bracket. Fink calls this phenomenology's "methodological schizophrenia" (Fink, cited in Overgaard 2015). My basic claim here is that choosing to wear one hat or another depends on the reasons for wearing that hat. This gives phenomenology an abductive aspect, both in terms of hypothesis generation (i.e., phenomenology of X promises to capture something missing in philosophy of Y), and in terms of justification (i.e., X is what has been revealed through this or that approach, which hangs together with Y, challenges Z, etc.). This can include metaphysical, methodological, and even scientific claims. In each of these cases, phenomenology can help to reveal the presuppositions of an inquiry and enable consideration of alternatives. If this is right, it seems that phenomenology is not just pure description without explanatory remit and ambition. Nor is phenomenology somehow pre-or post-argumentative, on account of these abductive inferences that are always implicitly at work and often explicit when the view is framed against an alternative theoretical and philosophical perspectives.
Elsewhere, I have argued that this general perspective accords with some crucial aspects of Merleau-Ponty's philosophy, even if not necessarily all of the remarks in the Preface that enigmatically introduce the method (Reynolds, 2017(Reynolds, , 2018. I think it is also plausible to interpret what Sartre offers in the totality of Being and nothingness as an inference to the best explanation for a metaphysical view. I will try to motivate the basic idea, albeit without doing any detailed scholarly work here. 6 First, Sartre discovers the phenomenological method around 1933 in the context of his existing criticisms of "digestive philosophy" (idealism) and reductive realism. These themes endure in his philosophy thereafter. The more strictly phenomenological moments of Being and nothingness subserve a particular metaphysical picture, and they play a role in its justification, but always in juxtaposition with other considerations, notably including the alleged insufficiencies of idealism and realism regarding the major themes of the text (Sartre, 1958). Sartre holds that we have a range of phenomena that a metaphysical view must be able to accommodate, or convincingly explain away, and these revolve around the experiences of negation, nothingness, and freedom. Sartre argues that this is not something that alternative views like idealism and realism can do. Hence, the Sartrean inference: either the counter-intuitive, non-naturalist metaphysics he propounds is correct, or negation is somehow rendered non-real, along with a range of human experiences that appear to depend on it. While Merleau-Ponty argues that there are neglected possibilities hidden by Sartre's antinomic thinking, Sartre's general strategy bears some structural similarities to Merleau-Ponty's own efforts to navigate the empiricism/intellectualism dialectic that we briefly considered above.
One might protest that there are parts of Being and nothingness or Phenomenology of perception that are properly phenomenological. In a relative sense, I agree: there are parts of these works that are more strictly descriptive of lived-experience. My claim however is that it is a relative matter, and that these descriptions are always contextually embedded, and have meaning in that context and with that background. They are always minimally argumentative, being couched in regard to broader considerations from philosophy and from the life-world that are never completely bracketed or put out of play. These argumentative and abductive aspects are present as the background and condition for the phenomenology. They frame the inquiries that take place, both in terms of where to begin but also where to end a phenomenological description and reflection, as well as the way in which one philosophically reflects and composes texts, seeking justification for a synthetic vision that possesses some minimal unity and direction 7 . There is always an abductive leap in terms of hypothesis generation and the methods felt to be appropriate for a given ambition. There is also a leap in terms of the synthetic justification for a given view that is being propounded. This more dialectical element is also at least minimally part of, rather than extraneous to, phenomenological description. To put this in terms imported from philosophy of science, the contexts of hypothesis formation (what to describe and 'why') and discovery (in this case description under the reduction, the 'how') are coimplicated in practice, although they are separable in principle. I will make a similar claim about transcendental phenomenology in what follows.

Transcendental phenomenology and abduction
So far I have focused on the aim of phenomenological description, noting that it is a special sort of description conducted in the light of a methodological technique (the phenomenological reduction). I have argued that this description is hermeneutically quite deep, involving abductive inference in terms of hypothesis generation and justification. An additional aspect of phenomenology, however, is its transcendental dimension, which is also typically argued not to involve an inference of the sorts characteristic of ordinary inductive reasoning, nor an IBE. After Husserl's "transcendental turn", commonly dated to around 1906/7, phenomenology becomes aligned with the transcendental search for essential conditions for types or kinds of experience. Although this is sometimes contested 8 , the idea of essential or necessary conditions involves a philosophical argument that might be considered in its logical form, even if we cannot reduce the whole question of "the transcendental" to that logical form. Regarding the argument form, the basic aim is to get from something subjector agent-involving to a claim about what must be the case for some phenomena to be as it is. This must involve ampliative or new knowledge in some way. Such knowledge might regard other concepts or structures that must be presupposed, or be about the nature of the world per se.
To get clearer about this, we can briefly consider Being and nothingness and Sartre's infamous key-hole example. One is not forced to accept his famous phenomenological description, but on Sartre's account we have a pre-moral experience of shame when we are 'caught' peering through the key-hole, which involves an abrupt pre-judicative transition from being given over to our environment and what we are perceiving, to apprehending ourselves as an object in the eyes of another. He argues that the existence of the other (qua ontological structure) is necessary for this experience to be as it is. However, we should note that this is part of an unfolding argument for a given metaphysical picture. This includes Sartre's dialectical engagement with idealism and realism, which is explicitly in the foreground of earlier material on "the reef of solipsism" that motivates his account of the Look. The success or otherwise of Sartre's transcendental argument here has been debated (cf. Sacks, 2005;Hyslop, 2008). My aim here is not to defend this argument, but simply to note that there are forms of reasoning and argument clearly embedded in it. I also want to emphasise 8 Some contended that it should be characterized as transcendental reflection rather than transcendental argument. As Crowell (1999) puts this point, "for Husserl the task is not to validate synthetic a priori judgments by means of transcendental arguments, but to clarify, by means of a kind of reflection, the sense of what shows up as real ('transcendent') in various modes of experience" (p. 32). This reminder that phenomenological claims are often about meaning or sense is well-taken and reaffirms the "nonargumentocentric" point. But the opposition between reflection and argument is a little exaggerated. Phenomenologists often maintain that various modes of experience exhibit a hierarchical structure, with certain conscious acts claimed to be more or less 'basic' to the extent that they presuppose other conscious acts. As such, they map the alleged possibility conditions for certain experiences, and an argument to this effect can be assembled (cf. Russell & Reynolds, 2011, pp. 301-2). If there is an opposition between transcendental reflection (or experience) and transcendental argument, there are nonetheless cases of both in phenomenology. I will show in the remainder of this paper that Husserl's argument about time-consciousness appears more classically argumentative, as does Sartre's arguments about intersubjectivity. that the putative conclusion of this argument is less deductive than it is suggestive and must be understood within the overall architectonic of the text. Sartre seems to think that this is not just a good explanation, but that it is the best and only viable explanation insofar as it gives "necessary and sufficient" conditions (1958, p. 250) 9 . But it is difficult to determine how Sartre establishes this, a difficulty that reflects an ongoing issue at the heart of transcendental arguments.
To further explicate this view, we can invoke a more naturalistic and mundane perspective. As I briefly noted in Sect. 1, the very fact that there is historical and synchronic diversity regarding descriptions of experience and claimed conditions within phenomenology is worth considering. On a formal level, we might wonder about the status of claims about necessary conditions that exhibit this idiosyncratic and apparently variable nature. A metaphysical conclusion is also commonly extracted through the claim that these alleged transcendental conditions are irreducible to the causal/ physical/empirical conditions provided by the kinds of naturalistic treatments briefly outlined above. But establishing this conclusion cannot simply be about the evidence of experience alone, within the terms of any 'reduction' (whether phenomenological or transcendental). Rather, the conclusion seems to depend on both empirical and conceptual work -that is, on analysis and explanation. For example, the conclusion depends on how we understand and define the causal, the physical, and the empirical, and the sorts of explanation they seek. Here, reductive explanations seem to be the main target, but they do not exhaust the field. Indeed, there is a risk that the claim of irreducibility concerning the transcendental dimension automatically follows from the restricted definition of naturalistic conditions (i.e., if the causal or physical order is understood as essentially linear and non-holistic, or as based on mere mechanisms, partes extra partes) 10 . But this restricted definition cannot be such that the conclusion of irreducibility trivially follows, since this would imply circularity. There are hence deep questions concerning whether transcendental conditions constitute a sui generis condition (or form of explanation), and concerning how one might block the argument that they (potentially) reduce to other sorts of empirical and philosophical treatments of conditions and necessity.
This is a version of the "uniqueness" condition that Stephan Körner and Barry Stroud showed is very difficult to establish in transcendental arguments (Körner, 1966;Stroud, 1967). Körner objects that a classical transcendental argument holds that: B obtains; A is a necessary condition of (possibility of) B, and therefore (A) But the proponent of this argument usually has not established that Z, for instance, is not also a necessary condition of (B) If there are other conditions that might equally necessitate B, then we cannot assert that the necessity in question is philosophically significant or unique. This is because it may or may not be compatible with a variety of other conditions. Unless we have gone through and ruled out all other potential conditions, including those given by other phenomenological analyses and those of a more empirical/causal/physical trajectory, we cannot claim to have gained significant ampliative knowledge and ruled out other sorts of explanation or approaches to the issue (non-phenomenological).
In this context it is helpful to also consider Husserl's well-known argument in The phenomenology of internal time consciousness (Husserl, 1991) about the necessity of consciousness having a particular temporal structure. This structure is supposed to have retentive and protentive dimensions, which incorporate the "just past", along with anticipations of the future, into any 'now' moment. Husserl argues this structure is necessary for us to have the experience of hearing a melody as we do (e.g., as connected and having a structure, rather than as a series of isolated notes or a cacophony). But claiming autonomy or privilege for this transcendental condition (i.e., as non-reducible to other kinds of explanation and other kinds of conditions) requires additional philosophical work. As Stroud (1967) argues, it is hard to see how that special or unique kind of condition and necessity (transcendental rather than empirical/causal/physical) might itself be established in strictly a priori fashion, without assuming idealism, and without assuming that a particular method is reliable and correct. The charge here is that we either beg the question and/or we just see that it is so. But it remains to be determined what we can say to those who do not 'just see it is so'. We cannot appeal to our own eidetic intuitions to settle the more argumentative or dialectical question about phenomenology. Neither can we appeal to the scope and significance of the ostensible "necessity" at stake in our phenomenological insights. As noted, this is contested even within the field of phenomenology.
We are left to wonder whether transcendental reasoning should be eliminated, or whether we might reconceive what the argument can deliver. Inkpin (2014) has argued against the transcendental dimension of phenomenology and for a version of "minimal phenomenology". Others have considered the possibility of maintaining a "weak" or "modest" role for transcendental arguments (e.g., Taylor 1995;Stern, 2007), but they have primarily remained committed to conceptual claims concerning necessity. These "modest" renderings often aim simply to leave aside any claim to get from how things appear "for us" to how things are. This involves backing away from some of the metaphysical claims present in Sartre's work, for example. There is, however, a middle option to which little attention is given (excepting in pragmatism). This middle option understands the transcendental as more akin to an abductive argument, and sometimes even an IBE, though these positions are not advanced in these terms. It means that strict necessity is given up, but one is still able to deploy phenomenological considerations to help advance metaphysical and epistemological arguments.
Abduction, along with IBEs that have been associated since Harman (1965), are a central feature of much scientific reasoning. They are also a feature of ordinary reasoning, arguably part of what we call "common sense", which involves inferences that are neither deductive nor inductive in the standard manner, but can still be intelligent and well-motivated. Abduction and IBE are used in Darwin, for example. They are also endorsed by Peirce's "naturalism" and in the context of his own reservations about any "armchair" or a priori conception of the transcendental project (Peirce, 1992; cf. Gava, 2014 for a reconciliationist move). There are two main senses of abduction: the broadly Peircean abduction (which associates abduction with hypoth-esis generation), and the contemporary sense of an IBE. Both are commonly held to be non-phenomenological, and particularly antithetical to transcendental phenomenology (Zahavi & Gallagher, 2008b, pp. 90 − 1).
Of course, it is true that an IBE does not usually concern itself with any detailed phenomenological description. Moreover, it does not generally bracket questions relating to the existence of some phenomena to better attend to the variegated structures of our experience itself. This is indicated by the fact that IBE is central to most arguments for scientific realism, whether concerning the epistemic credentials of the theories in question or the existence of particular entities that are not directly observable with unaided perception. With Putnam's "no miracles" argument, for example, the explanatory inference functions in cursory form: either we infer that the relevant sciences that support aeronautical engineering, aviation, etc., posit real entities and broadly truth-conducive theories, or we have no explanation at all for the fact that planes (generally) stay in the air (1975). Our only alternative explanation, so Putnam's story goes, is miracles (Putnam, 1975, p. 73). This forced choice is based on a rather abstract philosophical argument, but it uses a similar basic inference to Sartre in Being and nothingness, though Sartre's argument is augmented by detailed phenomenological descriptions of phenomena that are said to depend on negation. I have also argued that a more mundane and Peircean sort of abductive inference appears to be involved in phenomenological reasoning of a 'local' sort (where to begin and end the phenomenological description), and in motivating the turn to transcendental phenomenology, a background that is never entirely done away with in any presuppositionless phenomenological description. It is important to recognise, with Williamson (2016), that abduction is not just about empirically oriented philosophy and an inference to the best causal explanation, but has a more general role in theory construction with a wider scope 11 .
In this respect, it is useful to compare the formal structure of Peirce's conception of abduction to transcendental phenomenological argumentation, especially in order to begin to get a sense of what any such reconstruction in these terms might look like. An abductive argument, for Peirce, does not establish necessity. Rather, the argument has something like the following structure: ally or analytically. This is the basis for the claim that the arguments yield synthetic a priori knowledge. Any ampliative argument for new knowledge aims to show that for this or that phenomena to be as it is, this or that must also be the case. We might consider again Husserl's example of the experience of a melody. According to Husserl, our experience of the melody is integrated, with the tune heading somewhere and having an anticipated structure, even on first listening. Husserl argues that a condition for this is that time-consciousness has retentive and protentive aspects. This seems plausible, but we might question whether it gives us strict necessity, as well as its scope. It might not pertain to the conscious experience of octopuses, say, which lack the kind of integrated neuro-cognitive system that humans have. Husserl himself may not be especially interested in this extension of scope. He might instead be aiming to reveal necessary structures and correlations within experience. But even then it is unclear whether these structures and correlations track something that obtains for any and all auditory experience, or whether they track altered or psycho-pathological states. Here, the transcendental argument requires a check and balance of some kind if it is intended to quell sceptical concerns (Stern, 2007), which in this context could include anyone who does not accept the ostensible necessity in question -that is, whether they do not accept the argument I have presented, or even if they do not arrive at the same conclusion via their own phenomenological reflection. There are many ways to do this kind of a check and balance, which the method(s) of phenomenology can assist with, even if they are not strictly transcendental in their approach. One is to look fully into factual variation within human experience, within psychopathology for example, and attempt to do so while remaining as presuppositionless as possible. Many phenomenologists undertake this sort of work. Other directions might include the sorts of second-person techniques involved in micro-phenomenological interviews (see Klinke and Fernandez 2022;Höffken 2021).
But to return to Husserl, his argument appears to be on stronger ground in relation to the negative conclusion that time-consciousness must not be a series of concatenated instants, because on such a view the phenomena of experiencing a melody, as described, appears inexplicable. It remains possible, however, that we can develop theories of 'discrete nows', atomistic views, that do address becoming. Various presentist views in the metaphysics of time have attempted this, as have neuro-cognitive models, possibly including some construals of predictive processing 12 . These views may not be correct, but it is not clear that any transcendental phenomenological argument alone can rule them out. Efforts to do so tend to take the conclusion for granted, or they caricature the "discrete nows" views so that they cannot address becoming and dynamism. Even if we take Husserl to provide grounds for thinking that timeconsciousness is unlikely to be adequately explained through a model that involves a series of concatenated instants, there may not be reason to think that all reductive treatments are (or will be) inadequate. This is what requires various methods of examination and testing, both from the abductive arm-chair à la Williamson (2016) and potentially with more empirical resources.
In this respect, however, we can also frame Husserl's own argument about timeconsciousness abductively (and fallibly) without losing much of its significance. When we perform the phenomenological reduction (which we should continue to do, noting it is always incomplete and imperfectly achieved), we might find that some class of our experiences does not square with our standard, pre-phenomenological conception, perhaps because of common-sense assumptions, or a metaphysical or scientific picture over-determining our analysis. 13 We can take Ernst Mach's famous picture of his visual scene (Mach, 1900, p. 13) as an example. This picture seems to capture the fundamentals of our visual scene in the natural attitude, at least until we attempt to attend to it phenomenologically, returning to the "things themselves". From that basis we might offer an alternative, phenomenological description of the visual scene -one that should draw attention to ambiguity and imprecision around the nose and at the periphery. Some of the details of that description might also be challenged. But insofar as the phenomenologist does a good job of capturing that experience from a more "lived" and first-personal perspective, then an argument regarding the conditions of this experience can be developed abductively rather than necessarily, by following Peirce's argument. To return to Husserl and time-consciousness, the abductive rendering of the view would be that we do not establish the strict necessity of protention and retention for all conscious experience; rather, we posit that the experience of the melody becomes more comprehensible (less surprising, less like a miracle) on this model than on others.
This gives the idea some prima facie support. We might then seek to build this up in various ways. From a phenomenological perspective, support might be gained by using tools like imaginative variation, which varies aspects of the phenomena in question to see what persists and appears to be essential. Such support might be developed in more empirical directions, through micro-phenomenological interviews, as indicated above. It might also include efforts to show (dialectically) that this structure of time-consciousness is presupposed in other competing explanations, including any ostensibly adynamical and atemporal accounts. Both attempts to bolster the basic abductive argument are potentially valuable, but it might be that no philosopher can achieve sufficient imaginative variations, or sufficient reductio-style treatments of all reductive or atomistic explanations. As such, it is arguable that the scope of the claim should be weakened, and that transcendental reasoning comes into contact with other forms of reasoning, perhaps even giving us something like a transcendental pragmatism 14 . If we are committed to thinking A (and B appears to us as a condition for A), and if there are other reasons which might suggest that B has a prima facie warrant and is the best explanation currently available, then we are entitled to take the tran-scendental reasoning beyond its natural ken. Whether or not any resulting synoptic view is the best explanation might be contested, since establishing "the best" is an ongoing process rather than something that might be attained once and for all. However, resulting views will purport to provide a better explanation than at least some of the major alternatives.
There is no deduction or strict proof in this process. We might make fallacious leaps of the sort that Husserl warned against. For Husserl (1999), as soon as we engage in ordinary reflection on knowledge, we "fall into error and confusion. We get involved in patently untenable positions, even in contradictions. We are in constant danger of falling into skepticism" (p. 18). This is indeed a risk of the approach I am advocating. But it is also a risk that is characteristic of the philosophical enterprise writ large. For the reasons I have outlined in this paper, it is not clear how we might immunise phenomenological reasoning against this risk without assuming a conception of transcendental reflection and insight that is question-begging and has some continuing logical and epistemological problems. As such, I think the dialectical conception of phenomenology which I advocate is both reasonable and inevitable, despite it being difficult to reconcile with some of the famous programmatic statements about phenomenology. This conception involves an abductive reading of the transcendental phenomenological tradition, at least beyond fairly banal necessities (i.e., all perception is perspectival or has a figure/ground structure, an object cannot be simultaneously red and green all over, etc.,), and insofar as the proclaimed transcendental dimension is mobilised to do further philosophical work, which it almost invariably is.

Conclusions
The aim of this paper was to show that phenomenology is always at least partly concerned with arguments and explanations, despite how it seems. In-itself, this is an important contribution, since the idea that phenomenology is non-argumentative is widely used by critics of phenomenology. I argued on behalf of phenomenology against these critics, but also against some aspects of phenomenology's own selfconception, wherein the argumentative aspect is also downplayed or treated as a nonphenomenological intrusion. Although the connection between phenomenology and dialectical argument has been examined before, few philosophers have considered the relationship between phenomenology and abductive reasoning. Usually, both major forms of abductive reasoning (as concerns hypothesis generation and justification) are considered antithetical to phenomenology. I have established reason to doubt this view by showing that a plausible version of transcendental phenomenological argumentation inevitably intersects with (and sometimes collapses into) these other argument forms. Instead, phenomenology and abduction might be mutually dependent on each other (i.e., in motivating the turn to phenomenology, in helping one determine when to end a phenomenological inquiry, etc.). We have also seen that modest framings of transcendental reasoning are also similar in form to an abductive "hunch" or IBE, and using that interpretation of them helps to avoid some of the dif-ficulties that afflict transcendental arguments on a more arm-chair conception, and thus there is reason to pursue such a conception of the formal nature of the argument.
In this way, phenomenology has been rescued from the problem of ostrich epistemology, but it may be wondered at what cost? Do we now need to abandon phenomenology or use it only in rhetorical fashion? Let me be clear that I don't think that conclusion is entailed. Phenomenology on Merleau-Ponty's conception remains viable and valuable, and indeed it is perhaps also inescapable, as I have argued elsewhere (C.f. Reynolds 2018). In addition, I do not mean to deny that there are more or less strictly descriptive aspects of phenomenology, even if I have argued that the stage of phenomenological discovery remains minimally abductive, as is the stage of justification, which is an intrinsic part of philosophical texts and arguments, of course. The reduction is an important part of the philosophical tool-kit, calling us back to the evidences of experience and returning to the "how" of experience rather than the "what". But we must remain vigilant about this phenomenological "switch" of focus and any claims to it being wholly presuppositionless, by considering factual variations as well as imaginative ones. Is this properly transcendental phenomenology? Judgments here will differ. I think that an enumeration of the 'how' of experience assists with the transcendental task of clarifying the conditions of possibility of appearance, but it is not alone sufficient for this task, for the reasons given in Sect. 2. Whether such a view remains a form of transcendental phenomenology remains an open question, one which will likely continue to divide phenomenologists.