Why Husserl is a Moderate Foundationalist

Foundationalism and coherentism are two fundamentally opposed basic epistemological views about the structure of justiﬁcation. Interestingly enough, there is no consensus on how to interpret Husserl. While interpreting Husserl as a foundationalist was the standard view in early Husserl scholarship, things have changed considerably as prominent commentators like Christian Beyer, John Drummond, Dagﬁnn Føllesdal, and Dan Zahavi have challenged this foundationalist interpretation. These anti-foundationalist interpretations have again been challenged, for instance, by Walter Hopp and Christian Erhard. One might suspect that inconsistencies in Husserl’s writings are the simple reason for this disagreement. I shall argue, however, that the real question is not so much how to read Husserl, but how to deﬁne foundationalism and that there is overwhelming textual evidence that Husserl championed the most tenable version of foundationalism: a moderate foundationalism that allows for incorporating coherentist elements.

only be justified if it coheres with other beliefs. The most obvious difference between foundationalism and coherentism is that the later denies the possibility of non-inferentially justified beliefs, i.e., so-called basic or foundational beliefs. Let us introduce these different theories more formally.
Immediate Justification: Non-inferential justification is possible Foundationalism: Justified beliefs are either non-inferentially justified or ultimately dependent upon non-inferentially justified beliefs Coherentism denies that beliefs can be non-inferentially justified and holds that a belief can only be justified by coherence with other beliefs. Following Pryor, we can distinguish between pure and impure coherentism (cf. Pryor 2005, p. 185 f.).
Pure Coherentism: Beliefs gain justification only through coherence with other beliefs Impure Coherentism: Beliefs can be justified only if they cohere with other beliefs, but also non-beliefs can play a justificatorily significant role Impure coherentists allow that non-beliefs such as perceptual experiences can contribute to the overall justification of one's beliefs, but they deny that non-beliefs can justify beliefs on their own. Non-beliefs can only play a supporting role, while coherence with other beliefs remains a necessary condition. Pryor exemplifies: ''For instance, an impure coherentist might say that when it looks to you as if you have hands, and you have justification to believe that your visual experiences are reliable, those facts together can make you justified in believing that you have hands.'' (Pryor 2005, p. 186) Coherentism-both in its pure and impure version-denies Immediate Justification, but the denial of Immediate Justification does not imply coherentism. 2 Foundationalism, on the other hand, affirms Immediate Justification, but the affirmation of Immediate Justification does not imply foundationalism. 3 Hence, foundationalism and coherentism not only differ in their respective answers to the question of whether immediate justification is possible, but it is safe to say that this question lies at the very heart of the whole debate and we thus need to take a closer look at Immediate Justification. 4 Plausible candidates for basic beliefs, i.e., beliefs that are immediately justified, are introspective beliefs like ''I feel pain'' or ''I have thoughts,'' rational beliefs such as ''1 ? 1 = 2,'' ethical beliefs such as ''Torturing for fun is wrong,'' and empirical beliefs like ''The door in front of me is closed.'' It is important to note that proponents of Immediate Justification and foundationalism do not deny that entertaining such basic beliefs is only possible if you have certain faculties, concepts, empirical input, etc. beforehand. Perhaps all our beliefs depend on other beliefs for their existence, but if Immediate Justification is right, some beliefs do not depend on other beliefs for their justification. As James Pryor puts it, ''the fact that you have immediate justification to believe P does not entail that no other beliefs are required for you to be able to form or entertain the belief that P. Having the concepts involved in the belief that P may require believing certain other propositions; it does not follow that any justification you have to believe P must be mediated by those other propositions.'' (Pryor 2005, p. 183) In recent epistemology, a distinction is made between strong, moderate, and weak foundationalism. These types of foundationalism differ in their respective understandings of the epistemic status of basic beliefs. Strong foundationalism holds that basic beliefs are ''not just adequately justified, but also infallible, certain, indubitable, or incorrigible.'' (BonJour 1985, p. 26) BonJour complains that '' [m]ost historical discussions of foundationalism and even many quite recent ones, both pro and con, have focused almost exclusively on strong foundationalism.'' (BonJour 1985, p. 27) Moderate foundationalists, on the other hand, are only committed to the view that basic beliefs possess a degree of justification that makes them epistemically justified on their own (cf. BonJour 1985, p. 26). Basic beliefs are adequately justified and do not need further justification through coherence with other beliefs. 5 BonJour is completely correct in saying that ''an adequate consideration of foundationalism need concern itself with nothing stronger than moderate foundationalism.'' (BonJour 1985, p. 28) While virtually every current foundationalist would agree with this statement, even in contemporary debates opponents of foundationalism regularly attack foundationalism without differentiating between strong and moderate foundationalism. Quite often foundationalism is attacked for its alleged commitment to the possibility of infallible justification. This criticism, however, only affects strong foundationalism. We will return to this topic below.
Lastly, according to weak foundationalism, basic beliefs possess only a very low degree of epistemic justification on their own, a degree of justification insufficient by itself either to satisfy the adequate-justification condition for knowledge or to qualify them as acceptable justifying premises for further beliefs. Such beliefs are only ''initially credible,'' rather than fully justified. (BonJour 1985, p. 28) Weak foundationalism is a hybrid between foundationalism and coherentism, which is why it is sometimes called ''foundherentism'' (Haack 1993). In my opinion, moderate foundationalism is the most tenable version of foundationalism, and I will show that there is convincing textual evidence that Husserl needs to be interpreted as a moderate foundationalist. Concerning weak foundationalism, I want to briefly mention two important objections. First, it is quite radical to deny that there are basic beliefs which are adequately justified on their own. My beliefs that I am in pain right now or that it seems to me that my laptop is black, do not seem to require evidential support from other beliefs in order to be justified. To put it differently, the claim that coherence is a necessary condition for justification would appear to be too strong. Surely, this objection applies not only to weak foundationalism but especially to all versions of coherentism. However, according to weak foundationalism, no belief can be justified in the absence of coherence and many philosophers take this requirement to be far too strong (cf., e.g., Grundmann 2008, p. 324). To be sure, you do not have to be a foundationalist to reject this requirement. All proponents of Immediate Justification have to. 6 The second objection more specifically addresses weak foundationalism. Like all versions of foundationalism, weak foundationalism is committed to the view that a non-basic belief can only be justified if it is inferred from a basic belief. This means that weak foundationalism denies the coherentist claim that mere coherence with other beliefs can be sufficient for justification. The story weak foundationalism tells is this. There are basic beliefs. These basic beliefs have some degree of justification on their own, but this degree of justification cannot be sufficient to be adequately justified. Basic beliefs can be adequately justified only if they cohere with other beliefs. All justified non-basic beliefs are inferred from adequately justified basic beliefs. The problem is this. If coherence alone cannot be sufficient to render a belief adequately justified, then it is mysterious how coherence can be supposed to increase the evidential status of a basic belief from ''not adequately justified'' to ''adequately justified'' (for this kind of criticism cf. BonJour 1985, p. 29).
As we have seen, strong, moderate, and weak foundationalism differ in their views concerning the degree of justification basic beliefs possess.
Strong foundationalism: Basic beliefs must be infallible.
Moderate foundationalism: Basic beliefs can be fallible (but are adequately justified on their own).
Weak foundationalism: Basic beliefs have some degree of justification, but they are not adequately justified on their own. In order to be adequately justified, even basic beliefs need to cohere with other beliefs.
We see that only strong foundationalism is committed to the view that our basic beliefs are infallible. This demand for infallibility is now rightly viewed as ''overkill'' (Bonjour 1985, p. 27). The two defining claims of foundationalism are that there are basic beliefs and that all justified non-basic beliefs can be inferred from basic beliefs. There is no genuine need for infallibility. We have to keep in mind that the label ''basic'' only tells us something about the structural role of a belief; it does not reveal the degree of justification a belief has. As James Pryor puts it, ''the question whether your justification to believe P is mediate or immediate is a question about what kind of epistemic support you have for P. It is not a question about how much support you have: nothing in our definition requires immediately justified beliefs to be infallible or indefeasible.'' (Pryor 2005, p. 183) Foundationalism does not say that the difference between immediate and mediate justification is a difference of justificatory quality, and it might well be that certain inferred beliefs might have a higher degree of justification than certain basic beliefs.
Interestingly enough, even leading epistemologists sometimes fail to recognize that not foundationalism per se, but only strong foundationalism is committed to the view that basic beliefs cannot be overturned. In her great work, Considered Judgment, Catherine Elgin stresses the epistemological role of ''initially credible judgments.'' Elgin emphasizes that her view is not foundationalist and one of the reasons she gives is that ' '[u]nlike the justification that attaches to foundationalism's basic claims, initial tenability can be lost.'' (Elgin 1996, p. 110) Elgin's view might or might not be foundationalist, but it simply is not true that foundationalists insist that basic beliefs cannot lose their justification. 7 Only strong foundationalists do. This is why contemporary foundationalists often use the term prima facie justification. Basic beliefs are prima facie justified, justified in the absence of defeaters.

The Most Promising Version of Foundationalism
In current debates, strong foundationalism is only of historical interest. One reason for this downfall is that it is doubtful whether beliefs/justification can really be infallible. The main reason, however, is that it is very unlikely that all our justified non-basic beliefs are inferred from infallible beliefs. This is particularly obvious when we consider our empirical beliefs. For instance, I believe that due to gravity my mug of coffee will fall on the ground if I drop it. While most philosophers would agree that such a belief is justified, virtually no one would insist that it can be inferred from an infallible basis. Surely, it is important to spell out what is meant by ''inferred from basic beliefs.'' If ''inferred'' was supposed to mean ''logically deduced,'' all versions of foundationalism would be unfeasible. The claim that all my justified non-basic empirical beliefs can be logically deduced from basic beliefs seems highly implausible. Hence, it is important to note that only some proponents of the strong, traditional version of foundationalism use ''inference'' in such a strong manner. Virtually every current proponent of foundationalism uses inference in a much more liberal way, which encompasses not only deductive reasoning but also, for instance, inductive and abductive reasoning and reasoning by analogy. Richard Fumerton defines inferential reasoning in the following way: ''Principle of Inferential Justification: To be justified in believing P on the basis of E one must not only be (1) justified in believing E, but also (2) justified in believing that E makes probable P.'' (Fumerton 2010) Of course, different foundationalists differ in their respective answers to the question of which types of reasoning can fulfill this principle, but, importantly, there is no necessary restriction to deductive reasoning.
So far we have discussed two common misconceptions about foundationalism. Contrary to widespread belief, foundationalism is neither incompatible with fallibilism, nor are foundationalists committed to the claim that all justified nonbasic beliefs can be logically deduced from basic beliefs.
Another point that needs to be addressed is the epistemic role of coherence. Foundationalist theories of justification are not incompatible with the inclusion of coherentist elements: For their part, foundationalists do not typically deny the power of coherence to contribute to the overall epistemic status of a body of belief. They simply insist that coherence cannot do all the work on its own-there must be at least a modicum of intrinsic credibility or non-inferential warrant possessed by basic beliefs before coherence can have its amplifying effect. (Van Cleve 2005, 169) There is nothing in the definition of foundationalism that speaks against the view that beliefs can get additional justificatory support from coherence with other beliefs. This justificatory support from coherence can work in all directions. Nonbasic beliefs can acquire additional support from their coherence with other nonbasic beliefs or/and their coherence with basic beliefs. Likewise, basic beliefs can acquire additional support from their coherence with other basic beliefs and/or their coherence with non-basic beliefs.
It might appear strange that basic beliefs can obtain justificatory support from non-basic beliefs. However, there is nothing wrong with that. It is even possible that the same belief can be immediately and inferentially justified (cf. Pryor 2005, p. 183). Plausible candidates for such a kind of epistemic overdetermination are mathematical beliefs. If you understand the concepts involved, you might be immediately justified in believing that 1 ? 1 = 2, but it is also possible to prove this rigorously.
So how do strong, moderate, and weak foundationalism answer the question of whether basic beliefs can acquire additional justificatory support through coherence?
Strong foundationalism: Additional support is impossible. Basic beliefs already possess the highest degree of justification.
Moderate foundationalism: Additional support is possible.
Weak foundationalism: Additional support is necessary.
Again, it is interesting to see that even leading epistemologists who are at the center of the debate sometimes fail to differentiate between strong foundationalism and more moderate versions. According to Susan Haack, for instance, ''foundationalism requires a distinction of basic versus derived beliefs and an essentially onedirectional notion of evidential support,'' which means that ''derived beliefs can give no support to basic beliefs.'' (Haack 2000, p. 226 f.) Elgin claims, ''Foundationalism's basic beliefs, however, cannot augment their original level of justification. They supply warrant to, but do not derive it from, the structures they support.'' (Elgin 1996, p. 110) These statements, however, solely apply to strong foundationalism.
We have seen that foundationalism comes in many different flavors. This is even true for moderate foundationalism. Moderate foundationalists differ in what they mean by inferential reasoning and in the epistemic role they ascribe to coherence.
In what follows in this section, I shall define the version of foundationalism that I take to be the most tenable one. This is also the kind of foundationalism I ascribe to Husserl. According to this version of moderate foundationalism: (1) There is immediate justification, i.e., there are basic beliefs that are adequately justified on their own. (2) Immediate justification does not have to be infallible.
(3) One's justified non-basic beliefs are inferred from one's basic beliefs.
(4) Inferential reasoning is understood in a liberal sense that is not restricted to deductive reasoning. (5) Inferential reasoning is fallible. (6) Coherence plays an important epistemic role. (7) Beliefs (even basic beliefs) can gain justificatory support from coherence with other beliefs. (8) In some cases, a lack of coherence can reduce the degree of justification a belief has. (9) Justified beliefs (either immediately or mediately justified) can be defeated by other beliefs.
(1)-(3) are the defining features of moderate foundationalism. Every moderate foundationalist must affirm these three claims. (4) and (5) are not inherent in the definition of moderate foundationalism, but for obvious reasons moderate foundationalists should and de facto do subscribe to them.
(6) is the statement some moderate foundationalists might disagree with or deem dispensable. The role the search for coherence plays in certain fields of cognitionmost notably in the natural sciences-makes me think that (6) is indispensable and a crucial improvement over traditional versions of foundationalism.
( 7) and (8) are simply meant to specify how coherence can play an important role in foundationalist systems. Imagine that two of your beliefs-B1 and B2-possess the same degree of prima facie justification. Let us further assume that B1 and B2 turn out to be incompatible with each other. They cannot both be true. Further investigation reveals that B1 is coherent with many of your well-justified beliefs. B2 is not. In this case, it seems natural to assume that B1's coherence with other beliefs increases its degree of justification, while B2's lack of coherence diminishes its degree of justification.
(9) is a statement most philosophers subscribe to. It is placed at the end of this enumeration since some might regard such a case of defeat as an extreme case of incoherence.
This section was supposed to show what moderate foundationalism is and what its most plausible version looks like. In the next section, we shall discuss whether Husserl subscribes to such a version of moderate foundationalism.

Husserl's Moderate Foundationalism
There are a considerable number of central passages in Husserl's oeuvre which suggest that Husserl must be interpreted as a foundationalist. This, combined with the fact that Husserl likes to stress his affinities to Descartes, might mislead one to interpret Husserl as a strong Cartesian foundationalist. Føllesdal was right in 1988 when he called interpreting Husserl as a foundationalist the ''universally accepted view'' (Føllesdal 1988, p. 115). Since then, things have changed considerably. Besides Føllesdal, further influential commentators such as Drummond, Zahavi, and most recently Beyer have challenged the foundationalist interpretation. One reason for this dramatic shift in interpretation is the insight that Husserl cannot be viewed as a strong foundationalist. Husserl clearly does not hold that all our justified beliefs have to be based on infallible beliefs. Moreover, many passages reveal that for Husserl coherence can play an important epistemic role.
As we have seen, however, holding that epistemic justification can be fallible and appreciating the importance of coherence does not mean that one is obliged to coherentism. Interpreting Husserl as a coherentist simply because he is not a strong foundationalist means failure to grasp what foundationalism is. At the least, it means the failure to distinguish different forms of foundationalism. As we will see, Føllesdal is clearly guilty of this failure.
In this section, I will first show that Husserl must be interpreted as a moderate foundationalist and, finally, I will address the worries raised by Beyer, Drummond, Føllesdal, and Zahavi.

Husserl's Conception of Intuition
Central to Husserl's epistemology is his conception of ''intuition''. 8 The logical first step in addressing Husserl's conception of intuition is to have a closer look at his Sixth Logical Investigation, where he makes the crucial distinction between signitive and intuitive acts. Intuitiveness refers to a mode of givenness. The same object can be given signitively or intuitively (cf. Hua XIX/2, p. 556). If the object is given signitively, we are directed toward the object via one of its meanings; what is given to us is not the object in its actual presence but the object as something that only is meant. If the object is given intuitively, the object is presented as ''bodily present,'' is given in a ''fleshed out'' manner. Signitive acts are ''empty'' acts, while intuitive acts have a ''presentive'' character. In a nutshell, signitive acts are empty intentions and intuitions are full intentions. When I think, for instance, ''there is a book on my desk and its color is red,'' this act of thinking is a signitive act. When I look at my desk and see that there is a red book on it, this act of seeing is an intuitive act, i.e., an intuition. 9 Closely related to the concept of intuition is the concept of fulfillment (cf. Investigation 6, §8). Fulfillment takes place when a full intention matches a corresponding empty intention, i.e., when the object is intuitively given as it is emptily intended. Fulfilling acts are intuitive acts that fulfill empty intentions. The easiest way to get a grip on all this is by means of a concrete example: A very enlightening one is delivered by Walter Hopp (Hopp 2015, pp. 181-184). Assume you wonder whether your office door is open. You might think ''my door is open,'' you might strongly believe so, but thinking that p or believing that p are signitive acts and have no justificatory force. You are not epistemically justified in believing that p simply because you think or believe that p. 10 If you want to know whether your door is open, the best thing you can do is go and check. Looking at your door and seeing that it is open is an intuitive act. The object/content ''my door is open'' is given to you in a presentive way. This intuitive act fulfills your empty intention. First, you wondered whether your door was open. You intended this state of affairs emptily. Now, your seeing that the door is open is a full intention that fulfills your empty intention.
It is such intuitive acts that seem to be the source of epistemic justification. When you see that your door is open, you are perfectly justified in believing that your door is open. Also, I would like to point out that for Husserl intuition is not restricted to sensory perception.
Roughly, Husserl distinguishes four types of intuitions, straightforward sense experiences of physical objects, categorial intuitions of states of affairs, essential intuitions of essences, and introspective intuitions of one's own mental states. 11 For Husserl, not only mathematical but all phenomenologically relevant insights are (based on) essential intuitions and what essential intuitions share with all forms of intuitions is their specific phenomenal character, sometimes termed originary givenness or self-givenness. 12 Husserl writes: Of whatever sort intuition of something individual may be, whether it be adequate or inadequate, it can take the turn into seeing an essence; and this seeing, whether it be correspondingly adequate or correspondingly inadequate, has the characteristic of a presentive act.
[…] Seeing an essence is also precisely intuition, just as an eidetic object is precisely an object. […] Empirical intuition or, specifically, experience, is consciousness of an individual object; and as an intuitive consciousness it ''makes this object given,'' as perception it makes an individual object given originarily in the consciousness of seizing upon this object ''originarily,'' in its ''personal'' 10 In current epistemology, the claim that believing p provides prima facie justification for p goes by the label ''doxastic conservatism'' and is now championed by virtually no one. 11 I want to thank an anonymous referee of this journal for suggesting to include introspective intuitions. For Husserl's distinction between different types of intuitions cf., e.g., Erhard (2012, p. 52), Kidd (2014), andSmith (2003, p. 29). 12 Strictly speaking, as we shall see in the next subsection, I should not say that all intuitions are justification-conferring acts that posses the character of self-givenness, since Husserl later uses the term ''originary presentive intuitions'' to distinguish such intuitions from intuitive acts such as imaginative acts that do not possess these characteristics. For the sake of simplicity, whenever I use the term ''intuition'' what I mean is ''originary presentive intuition'' such as sensory or essential intuition. selfhood. In quite the same manner intuition of an essence is consciousness of something, an ''object,'' a Something to which the intuitional regard is directed and which is ''itself given'' in the intuition. (Husserl 1983, p. 9f.; Hua III/1, p. 14f.) When I merely think ''the door in front of me is closed,'' this is a signitive act. This act of thinking has no justificatory force at all. But when I look at the door and see that it is closed, this act of seeing is in an intuitive act and is a source of immediate justification. Likewise, when I learn at school that 5*5 = 25, I might be justified in believing so since my teacher is a reliable source of mathematical knowledge. However, Husserl would insist that there is a huge difference between being able to think or being able to tell that 5*5 = 25 and having the essential intuition that 5*5 = 25 (cf. Hua XXX, pp. 325-326). 13 Only by intuiting am I immediately justified in believing that 5*5 = 25. 14 Moreover, essential intuitions have the unique characteristic of not only revealing that something is the case, but also why it necessarily must be so.

Intuition as a Source of Immediate Yet Fallible Justification
The foregoing should be sufficient to elucidate what Husserl means by ''intuition.'' I now wish to discuss the epistemic role intuition plays in Husserl's philosophical system. The significance of intuition is most prominently demonstrated in Husserl's famous principle of all principles (Ideas I, §24): No conceivable theory can make us err with respect to the principle of principles: that every originary presentive intuition is a legitimizing source of cognition, that everything originarily (so to speak, in its ''personal'' actuality) 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 1983, p. 44) Here Husserl unambiguously states that ''every originary presentive intuition'' is a source of justification. As noted above in footnote 12, Husserl uses the qualification ''originary presentive'' in order to highlight the presentive phenomenal character of such acts and to distinguish them from imaginative acts. For the sake of simplicity, I use the terms ''intuition'' and ''originary presentive intuition'' synonymously.
One might note that Husserl's principle of all principles seems to be stronger than interpreted by me. I said this passage reveals that every intuition is a source of justification. In fact, however, Husserl states that every intuition is a source of cognition (''Erkenntnis''). If by ''cognition'' Husserl means what we now 13 Similarly, in Ideas I Husserl states: ''We can assert 'blindly' that two plus one is equal to one plus two; but we can also make the same judgment in the manner peculiar to intellectual seeing.'' (Husserl 1983, p. 327; Hua III/1, p. 315). 14 As has been indicated here, there are different types of intuition and these different types of intuition correspond to different types of objects. This means, ''the precise nature of the intuitive evidence at issue varies in dependence on the respective types of objects or states of affairs that the relating acts are directed towards.'' (Rinofner-Kreidl 2014, p. 37) Sensory perceptions and mathematical intuitions are only the most prominent examples of intuitive acts. understand by ''knowledge,'' then his principle is too strong. Knowledge is factive. If p is known, p is true. If every intuition was a source of knowledge, intuition must be factive/veridical too. As intuitive acts are non-factive but can include illusions and hallucinations, not every intuition can be a source of knowledge. 15 In this sense, Hopp says that the first part of Husserl's principle of all principles ''appears to be false.'' (Hopp 2015, p. 184) Hence, the principle of all principles might be the passage in which the importance of intuition is most decisively stated, but not so in the most accurate way.
Perhaps the passage of Ideas I that best encapsulates Husserl's main epistemological principles with respect to intuition is the following one: Immediate ''seeing'', not merely sensuous, experiential seeing, but seeing in the universal sense as an originally presentive consciousness of any kind whatever, is the ultimate legitimizing source of all rational assertions. This source has its legitimizing function only because, and to the extent that, it is an originally presentive source.
[…] Moreover, as may be added here to prevent possible misinterpretations, that does not exclude the possibility that, under some circumstances, one seeing conflicts with another and likewise that one legitimate assertion conflicts with another. For that, perhaps, no more implies that seeing is not a legitimizing basis than the outweighing of one force by another signifies that the outweighed force is not a force. It does say, however, that perhaps in a certain category of intuitions (and that is the case precisely with sensuously experiencing intuitions) seeing is, according to its essence, ''imperfect,'' that of essential necessity it can become strengthened or weakened, that consequently an assertion having an immediate, and therefore genuine, legitimizing ground in experience nevertheless may have to be abandoned in the further course of experience because of a counter legitimacy outweighing and annulling it. (Husserl 1983, p. 36f.; Hua III/1, p. 43) From the context it is clear that ''immediate 'seeing''' is synonymous with intuition. This passage reveals that for Husserl intuition is ''the ultimate legitimizing source of all rational assertion.'' The emphasis here is on ''ultimate.'' From the principle of all principles we know that Husserl declares every intuition to be a source of immediate justification. Here we can see that intuition is the ultimate source of justification.
In the first section we defined, Foundationalism: Justified beliefs are either non-inferentially justified or ultimately dependent upon non-inferentially justified beliefs.
This means one is a foundationalist if and only if one subscribes to the following claims.
F1: There is immediate justification. (=There are basic beliefs.) F2: All justified non-basic beliefs are inferred from basic beliefs.
Accordingly, Husserl qualifies as a foundationalist, since he holds that intuitions are immediate justifiers and the ultimate source of justification. The question to be answered is what kind of foundationalist he is. He obviously is no weak foundationalist. Weak foundationalism holds that coherence is necessary for justification, which is at odds with Husserl's conviction that intuitions are an autonomous source of justification. The passage quoted above suggests that Husserl is a moderate foundationalist, since he states that ''an assertion having an immediate, and therefore genuine, legitimizing ground in experience nevertheless may have to be abandoned in the further course of experience because of a counter legitimacy outweighing and annulling it.'' Thus, Husserl is not committed to the view that (immediate) justification is infallible, which means he is not a strong but a moderate foundationalist. One remaining question is whether all kinds of intuition are fallible. Husserl distinguishes between inadequate and adequate/apodictic evidence. Until the Cartesian Meditations (1931) he uses adequate and apodictic evidence synonymously. Within the Cartesian Meditations Husserl distinguishes apodictic from adequate evidence, and consequently makes a threefold distinction between inadequate, adequate, and apodictic evidence.
The terms ''inadequate,'' ''adequate,'' and ''apodictic'' need to be interpreted phenomenologically. This means these terms denote how an object is presented or given within experience. Physical objects, for instance, can only be given inadequately. There are always aspects a physical object has that are not presented to you in your experience of this object. In Husserl's terminology, physical objects are, necessarily, given in adumbrations.
The first thing to note is that justification is not limited to adequate or apodictic givenness (cf. Hopp 2015, p. 182). Seeing physical objects can only deliver or amount to inadequate evidence, but we have already discussed that for Husserl such intuitions are prime examples of justifiers. One might wonder, however, whether apodictic evidence distinguishes itself by being infallible. In his Formal and Transcendental Logic Husserl is quite clear that this is not the case: The possibility of deception is inherent in the evidence of experience and does not annul either its fundamental character or its effect […] Even an ostensibly apodictic evidence can become disclosed as deception and, in that event, presupposes a similar evidence by which it is ''shattered''. (Husserl 1969, p. 156; Hua XVII, p. 164) 16 16 For the sake of accuracy, it should be mentioned that this passage allows for two readings. According to the first one, apodictic evidence is fallible. This is the reading I advocate. According to the second reading, only infallible evidence can really be apodictic while those evidences that happen to be false can only be ostensibly apodictic. Hence, this second reading implies the infallibility of (really) apodictic evidence. I prefer the first reading as I believe that evidence must be interpreted phenomenologically. Different types of evidence denote how their contents/objects are presented to us. Accordingly, an evidence that p is apodictic if it presents p as apodictic, i.e., if p is experienced as being absolutely indubitable. Such an interpretation is in accordance with Drummond (2007, p. 37f.) and Soffer (1991, p. 132f.). For an extensive and highly valuable discussion of this passage cf. Heffernan (2009). Enlightening remarks concerning the fallibility of a priori insights can be found in BonJour (1998, § §4.4, 4.5).
In this context it is important to note that Husserl even allows that in some cases empirical evidence can successfully cast into doubt justification gained by essential intuitions (cf. Hua XXV, p. 248). This not only reveals the fallibility of apodictic evidence but foreshadows that Husserl's moderate foundationalism acknowledges the epistemic significance of coherence.

Husserl's Version of Moderate Foundationalism
Having clarified that Husserl is a moderate foundationalist, let us have a look at the version of moderate foundationalism we deemed to be the most promising one.
Moderate foundationalism in its most sophisticated form: (1) There is immediate justification, i.e., there are basic beliefs that are adequately justified on their own. (2) Immediate justification does not have to be infallible.
(3) One's justified non-basic beliefs are inferred from one's basic beliefs.
(4) Inferential reasoning is understood in a liberal sense that is not restricted to deductive reasoning. (5) Inferential reasoning is fallible. (6) Coherence plays an important epistemic role. (7) Beliefs (even basic beliefs) can gain justificatory support from coherence with other beliefs. (8) In some cases, a lack of coherence can reduce the degree of justification a belief has. (9) Justified beliefs (either immediately or mediately justified) can be defeated by other beliefs.
So far, we have seen that Husserl's system is in agreement with the first three and the last claim. Husserl's version of moderate foundationalism (so far): (1*) Intuition is a source of immediate justification.
(3*) Intuition is the ultimate source of justification. … (9*) Evidence can be shattered by other evidence.
Concerning inferential reasoning, one of the most decisive passages I know is to be found near the end of Ideas I.
As we know, all mediate grounding leads back to immediate grounding. With respect to all object-provinces and positings related to them, the primal source of all legitimacy lies in immediate evidence, more narrowly delimited, in originary evidence, or in the originary givenness motivating it. But in different ways one can indirectly draw upon this source, deriving from it the rational value of a positing which, in itself, is not an evidence, or if the positing is immediate, confirming and corroborating it. (Husserl 1983, p. 338;Hua III/1, p. 326) This passage is remarkable in many ways. First, it clearly shows Husserl's foundationalist attitude. Originary evidence is immediate evidence and in ''all object-provinces and positings related to them,'' (i.e., in all fields of possible knowledge) ''all mediate grounding leads back to immediate grounding,'' which means that all mediate justification has to be based on immediate justification. Second, this passage reveals that for Husserl there are ''different ways'' in which ''one can indirectly draw upon this [immediate] source.'' This means that Husserl also agrees with claim (5). Inferential reasoning is not limited to deductive reasoning, as there are different forms of inferential reasoning. Husserl correctly stresses that in mathematics inferential reasoning is deductive reasoning (cf., e.g., Hua III/1, pp. 22, 152f.). Natural sciences, by contrast, proceed inductively (cf., e.g., Hua III/1, p. 171; Hua VII, pp. 88, 102; Hua VIII, p. 240). Husserl holds that ''inductive empirical inquiry'' is only possible ''by virtue of an antecedent inquiry into essences.'' (Husserl 1969, p. 246;Hua XVII, p 252f.) This is a rather bold claim. What is of importance here, however, is that deductive reasoning, as well as inductive reasoning, are for Husserl acceptable forms of inferential reasoning as can be seen when Husserl poses the following question: ''Now, we ask, what about the truth of mediate inference, regardless of whether or not they are deductive or inductive?'' (Husserl 1983, p. 38;Hua III/1, p. 44) Enlightening is Husserl's remark that [corrected] experience-either as communal experience and reciprocal correction or as one's own personal experience and self-correction-does not change the relativity of experience; even as communal experience it is relative, and thus all descriptive assertions are necessarily relative, and all conceivable inferences, deductive or inductive, are relative. (Husserl 1970, p. 336;Hua VI, p. 270) Here Husserl stresses the corrigibility of experience and he leaves no doubt that inferential reasoning includes deductive as well as inductive reasoning. Perhaps even more importantly, this passage tells us that for Husserl even deductive reasoning is fallible. 17 This means that Husserl's moderate foundationalism also fulfills claim (5). In phenomenological inquiry, by the way, inferential reasoning has, according to Husserl, only a ''methodic function'': According to what has been stated, deductive theorizings are excluded from phenomenology. Mediate inferences are not exactly denied to it; but, since all its cognitions ought to be descriptive, purely befitting the immanental sphere, inferences, non-intuitive modes of procedure of any kind, only have the methodic function of leading us to the matters in question upon which a subsequent direct seeing of essences must make given. Analogies which emerge may suggest presumed likelihoods about concatenations of essences prior to actual intuition, and conclusions may be drawn from them; but ultimately an actual seeing of the concatenations of essences must redeem the presumed likelihoods. As long as that has not occurred, we have no phenomenological result. (Husserl 1983, p. 169;Hua III/1, p. 157f.) This passage suggests that for Husserl reasoning by analogy is a further acceptable form of inferential reasoning (even if such inferential reasoning only has a methodic function in phenomenology.) This should suffice to show that Husserl subscribes to (4) and (5).
Let us now turn to the important question of what role coherence plays in Husserl's system. According to Føllesdal, Husserl's mature work ''developed more and more in the direction of a quite extreme holism.'' (Føllesdal 1988, p. 107) With respect to the role of coherence, Føllesdal makes clear that ''Husserl again and again emphasizes the importance of coherence for validity and for truth.'' (Føllesdal 1988, p. 121) The textual evidence Føllesdal presents is a passage from Erste Philosophie which Føllesdal translates as A judgment-unity penetrates all the individual judgments.
[…] they have a unity which builds itself up in the progression of judgment, tying together judgment-sense with judgment-sense. This unity confers on all of them an intrinsic, interrelated validity. In this way the multiple statements in a treatise have a comprehensive judgment-unity, and so has in its way every theory and every entire science. (Hua VII, p. 19 translated by Føllesdal in Føllesdal 1988, p. 121) Here two things need to be pointed out. First, this statement is no commitment to extreme holism, it is not even a commitment to weak coherentism as it does not indicate that coherence or ''unity'' is necessary for justification or ''validity''. Secondly, the crucial passage, ''This unity confers on all of them an intrinsic, interrelated validity'' might be better translated as ''This confers to all of them the unity of an internally interdependent validity''. 18 This passage is not so much about epistemic justification, as about how single beliefs are related to each other within a coherent belief-system. This might reveal that coherence plays a significant role for Husserl, but is fully compatible with foundationalism.
According to Beyer, epistemic justification in Husserl is holistic since Husserl holds that coherence can produce knowledge (cf. Beyer 2012, p. 36). The textual evidence he delivers is a passage from the second part of Erste Philosophie, where Husserl states that the possibility of the world being a mere illusion ''is a mere empty possibility that is at odds with all empirical evidence and the full force of its unity. Nothing speaks in favor of this empty possibility''. 19  Again, this statement is fully compatible with foundationalism. It suggests that for Husserl beliefs can gain justificatory support from coherence with other beliefs and that a lack of coherence can reduce the degree of justification a belief has. (These are the claims (7) and (8) of our preferred version of moderate foundationalism.) This passage, however, does not imply that coherence is necessary or sufficient for justification.
Another strong evidence for our claim that Husserl acknowledges the epistemic force of coherence and seeks to incorporate coherentist elements in his version of moderate foundationalism is his conception of verification (''Bewährung''), which is supposed to point out that our beliefs are revisable and increase in their justification by being tested constantly (cf., e.g., Hua I, pp. 52f., 93; Hua XXIV, p. 345f.). Especially important is a passage at the end of Crisis, where Husserl stresses that verification can be reached by the unity (''Einstimmigkeit'') of experience (cf. Hua VI, p. 270). Such a unity of experience, surely, ''does not prevent the possibility of illusion'' (Husserl 1970, p. 335;Hua VI, p. 270), but Husserl leaves no doubt about the epistemic importance of such a verification by unity of experience.
Thus, Husserl refers to the unity of our empirical evidence when he rejects skepticism about the external world, and he highlights the importance of verifying our experiences by their unity. However, one might object that this is still insufficient textual evidence for the claim that Husserl subscribes to (7) Beliefs (even basic beliefs) can gain justificatory support from coherence with other beliefs.
One can find such textual evidence in Husserl's discussion of the justificatory force of memory in §141 of Ideas I: In a certain mode, any clear memory has original, immediate legitimacy: considered in and for itself, it ''weighs'' something, no matter how much or how little; it has a ''weight.'' But it only has a relative and imperfect legitimacy […] With every advance from memory to memory into the concatenation of memories being made distinct, memory is confirmed. To a certain extent the confirmation is a reciprocal one, the weights of the memories are functionally dependent on one another; each memory in the concatenation has a growing force with the amplification of the concatenation-a force greater than they would have in a narrower concatenation or in isolation. (Husserl 1983, p. 338;Hua III/1, p. 326f.) According to Husserl, there is a reciprocal confirmation of our memories and the more memories there are in such a concatenation of memories the more justificatory force these memories have. 20 This shows quite clearly that Husserl acknowledges the significance of coherence and subscribes to the claim that beliefs can gain justificatory support from coherence with other beliefs. To remove even the slightest doubt that Husserl endorses (7), take a look at the following passage: Moreover, if we proceed further in the context of perceptions and memories, then it becomes clear that the empirical weight increases with the comprehensiveness of the empirical context into which the individual experience concerned fits, i.e., fits without contradiction, therefore, becomes increasingly entwined with its neighbors in the sense of confirmatory agreement. (Husserl 2008, p. 343;Hua XXIV, p. 347) This shows that for Husserl a single experience's justificatory force is increased if it is in ''confirmatory agreement'' with its fellow experiences. As pointed out by Erhard, Husserl's wording in this quote not only shows that he acknowledges the importance of coherence, but also that he does so in a foundationalist manner. Coherence increases justification, it does not produce justification (Erhard 2012, p. 57, fn 27).
Surely, not all details of Husserl's moderate foundationlism are fully elaborated and there are some flaws and shortcomings. 21 However, this should only motivate current phenomenologists to further improve Husserl's epistemological system. Let us have an updated look at this system: Husserl's version of moderate foundationalism (updated): (1*) Intuition is a source of immediate justification.
(3*) Intuition is the ultimate source of justification.
(4) Inferential reasoning is understood in a liberal sense that is not restricted to deductive reasoning.
(8) In some cases, a lack of coherence can reduce the degree of justification a belief has. (9*) Evidence can be shattered by other evidence.
To be sure, this enumeration is not complete. Most importantly, there remains the question of what it is, according to Husserl, that makes intuitions justifiers. What gives intuitions their justificatory force? In my opinion, Husserl endorses the view that originary presentive intuitions are justifiers precisely due to their originary presentive character, i.e., their distinctive, justification-conferring phenomenal 21 Apart from terminological adaptations and clarifications, there is the need of elaborating the role of coherence in more detail, especially concerning the relationship between basic and non-basic beliefs. Husserl claims that evidence can be shattered by new evidence, but is this supposed to mean that beliefs based on originary intuitions can only be overruled by new originary intuitions? What about, for instance, justification by testimony? Current phenomenology has to engage in contemporary analytical debates and deal with recent topics in order to address ''new'' problems like the evidential status of testimony. Between phenomenology and current analytic epistemology there is a great potential for rich and mutually beneficial synergies. For me, it is beyond doubt that current debates in analytic philosophy could profit significantly from adopting elements of Husserl's phenomenology, first and foremost his conceptions of intuition and evidence.
character. 22 In current analytic epistemology, there is a movement that shares this basic idea that certain experiences are justifiers due to their distinctive phenomenal character and some proponents are aware of this close affinity to Husserl (cf. Chudnoff 2013 andChurch 2013). This, however, is another story.

An Analysis of, and Response to, Opposing Interpretations
After having established that Husserl has to be interpreted as a moderate foundationalist, we turn to the question of why Husserl's phenomenology is often considered incompatible with foundationalism. Especially common are prejudices of the forms: P1: In foundational systems, (immediate) justification must be infallible. P2: In foundational systems, inferential reasoning is restricted to deductive reasoning. P3: In foundational systems, coherence can play no important epistemic role.
As we have seen, Føllesdal (1988) seems to be committed to all three of these prejudices. Most obviously, he champions P1. This is why Hopp states: I don't discuss his arguments here because he defines foundationalism as the view that ''one can reach absolute certainty, at least concerning some matters, and also that […] it [is] a main task of philosophy to attain such certainty.'' (p. 107) Though Føllesdal presents a strong case that Husserl did not endorse such a position, infallibilism of this sort is not necessary for foundationalism. (Hopp 2008, p. 214, note 9) Drummond seems to be guilty of P2. To be sure, Drummond's works on the foundationalism debate are outstanding and highly instructive, and in Drummond 1991 he makes the crucial distinction between strong and moderate foundationalism. 23 However, when Drummond argues against the plausibility of moderate foundationalism, he complains that ''it is difficult to see how a non-propositional content would logically justify a propositional content.'' (Drummond 1991, p. 57) This complaint plays a crucial role in his rejection of foundationalism. Hopp refers to the same passage when he writes: And Drummond appears to believe that, if Epistemological Foundationalism is true, then the justificatory relations between basic and non-basic mental states must be logical. Drummond is surely right that, historically, most foundationalists, and the majority of coherentists, have construed epistemic justification as parasitical upon logical relations among the contents of the 22 Remember that in Ideas I Husserl tells us that immediate seeing, i.e., originary presentive intuition, ''has its legitimizing function only because, and to the extent that, it is an originally presentive source.'' (Husserl 1983, p. 36;Hua III/1, p. 43) Cf. also Hua XXIV, p. 347, where Husserl calls this justificationconferring phenomenal character ''consciousness of givenness.''. 23 One of his essential achievements is to show that Husserl's talk of apodictic and indubitable evidence does not commit Husserl to strong foundationalism. For Husserl, indubitability does not imply infallibility (Drummond 1991, p. 65). mental states involved. But that is not something to which the Epistemological Foundationalist must be committed, nor does it follow from Drummond's own definition. (Hopp 2008, p. 198f.) It is not entirely clear to me whether Drummond here wants to say that foundationalism is committed to the claim that inferential reasoning has to be deductive reasoning or to the claim that all justifiers have to be propositional. It is important to note that foundationalism is committed to neither of these claims.
In his Historical Dictionary of Husserl's Philosophy Drummond defines foundationalism in the following way: Foundationalism is a philosophical position that claims that there are bedrock or fundamental truths that underlie the justification of all other truths. These truths concern the world and all other truths about the world are logically derived [my emphasis] from them, or they concern the nature of cognition and it is by reference to them that all other knowledge is justified. Once these truths are known, in other words, it is possible to determine the truth or falsity of all beliefs. (Drummond 2007, p. 81) This definition is highly problematic as it-with respect to beliefs concerning the world-straightforwardly implies P2. As we have seen, moderate foundationalism is not committed to the view that all beliefs concerning the world are either foundational or logically derived from foundational beliefs.
Zahavi argues: [T]here is, however, something very misleading about calling Husserl a foundationalist, at least if the term is used in its traditional epistemological sense. As Husserl himself observes in Formale und Transzendentale Logik, the very attempt to establish a science which is based exclusively on absolutely certain truths is one that ultimately involves a misunderstanding of the very nature of science (Hua 17/169). As it was pointed out in Part I, Husserl's own concept of evidence by no means excludes errors or subsequent corrections (cf. also p. 138). (Zahavi 2003, p. 67) Furthermore Zahavi points out that for Husserl ''[p]hilosophy as a science based on ultimate justification is an idea which can only be realized in an infinite historical process (Hua 8/186, 6/439)'' and that ''Husserl explicitly distances himself from the axiomatic and deductive ideal of method that rationalistic foundationalism has normally been committed to (Hua I/63).'' (Zahavi 2003, p. 67) Zahavi is right that Husserl is neither committed to infallibilism nor to the ideal of axiomatic justification. But neither is foundationalism and there is nothing misleading about calling Husserl a moderate foundationalist.
Recently, Beyer has stated that Husserl is not a foundationalist (Beyer 2012, p. 42). 24 What leads him to this conclusion? Which Husserlian doctrine is deemed to be incompatible with foundationalism? Unfortunately, Beyer is not very precise on this. His line of reasoning, however, suggests that he is committed to P3. When arguing against interpreting Husserl as a foundationalist, he cites passages in which Husserl stresses the importance of coherence. We have discussed above such a passage quoted by Beyer, namely Hua VIII, p. 67, and came to the conclusion that such passages are not counterexamples to our claim that Husserl is a moderate foundationalist.
Beyer also draws attention to Husserl's discovery of the horizonal structure of experience. Experiences usually go beyond what is directly given. Physical objects, for instance, always appear perspectivally, they are necessarily given in adumbrations. They can never be given in their totality as there are always aspects and features that are currently not perceived by the same agent. However, there is always something co-given in our experiences. There are aspects, features, and even possibilities that are co-intended. When I look at the red mug of coffee in front of me, I am only directly aware of its front side. My intuition, however, is not solely directed toward the front side of the mug but towards the object in its entirety. I only see the front side, but I am aware of the fact that the whole mug is red and of what it looks like from different angles. I have expectations how the appearance of the mug changes when I come closer, how it feels when I grab it, and how its content tastes. I am aware that I should grab it carefully, as it might shatter if it were dropped.
Our current perceptions are embedded in a manifold of possible ''perceptions that we could have, if we actively directed the course of perception otherwise: if, for example, we turned our eyes that way instead of this, or if we were to step forward or to one side, and so forth'' (Husserl 1960, p. 44). Intending, for Husserl, is always an ''intending-beyond-itself'' (Husserl 1960, p. 46). For Beyer, Husserl's emphasis on the fact that our current experiences are commonly embedded in a manifold of coherent co-intentions indicates Husserl's coherentist account of justification (Beyer 2012, p. 42). I disagree with this verdict and line of reasoning. In my opinion, Husserl's conception of horizonal intentionality tells us something about the structure of experience, not about the structure of justification. Even if Husserl claimed that experiences only have justificatory force due to their horizonal structure, this would not commit him to coherentism. Let us elaborate this in more detail. 25 Coherentism is the view that beliefs are justified only if they cohere with other beliefs. Note that this is true for both pure and impure coherentism. Impure coherentism allows that non-beliefs such as perceptual experiences can contribute to the overall justification of one's beliefs, but coherence with other beliefs is still a necessary condition for justification. In other words, every kind of coherentism denies that experiencing p can be sufficient for justifiably believing that p. Thus, the question is: Does Husserl believe that experiences can be an autonomous source of justification? Can a belief based on an experience be justified without epistemic support from other beliefs? Husserl clearly gives an affirmative answer. ''Simple perception grounds a perceptual judgment. I see it, that's how it is. Why? I see it.'' (Husserl XXIV,p. 345) The point is that an experience or an experience's presentive, justification-conferring phenomenal character might depend for their existence on factors such as other experiences in the past or being embedded in a manifold of co-intentions, but this does not imply that a belief based on an experience depends for its justification on other beliefs. My experiencing the mug in front of me can immediately justify my belief that there is a mug in front of me even if this experience is embedded in or consists of a coherent manifold of (co-) intentions.
However, many commentators seem sympathetic to the following line of reasoning: 1. Every perceptual experience depends on (cohering) horizonal contents for its presentive character. 2. Every perceptual experience depends for its epistemic power on its presentive character. 3. Therefore, every perceptual experience depends on (cohering) horizonal contents for its epistemic power. 4. If an experience depends on horizonal contents for its epistemic power, then some version of coherentism is true. 5. Therefore, some version of coherentism is true. 26 The problem is with premise 4. Experiences may depend for their justificatory force on many factors, but this does not imply that they cannot be a source of immediate justification. Experiences can even depend for their justificatory force on other past experiences and still be a source of immediate justification. 27 This can be seen by considering cases of perceptual learning. Perceptual learning ''refers to an increase in the ability to extract information from the environment, as a result of experience and practice with stimulation coming from it'' (Gibson 1969, p. 3). Say you are looking at a piece of paper with two lines that slightly differ in length. At time t 1 you are unable to spot the difference in length. It visually seems to you that both lines have the same length. After some practice, the content of your experience has changed and at time t 2 you are able to spot the difference. It visually seems to you that the lines differ in length. Plausibly, at t 2 you are immediately justified in believing that the lines differ in length. You are, simply because you can see it.
Thus, in some sense, your experience depends for its justificatory force on other experiences, but this does not mean that a belief based on this experience depends for its justification on other beliefs. 28 Concerning the above argument, we can respond: The question is not whether your experience depends on certain factors for its presentive character. The question is: If your experience has a presentive 26 This argument was suggested to me by an anonymous referee of this journal. 27 Here dependence is used in the following sense: A depends for B on C if without the (prior) existence of C, A would not have the feature B. 28 In this context, consider also Husserl's conceptions of eidetic variation and essential intuition. A belief justified by an essential intuition is epistemically independent from other beliefs, i.e., immediately justified. However, eidetic variation which results in having an essential intuition is possible only in virtue of underlying mental states intentionally directed towards concrete objects or states of affairs. character, is it a source of immediate justification? Husserl says so. Because he says so, he is no coherentist. 29

Conclusion
Recently, several authors have rightfully stressed that Husserl advocates a conception of epistemic justification that acknowledges the fallibility of our justified beliefs and can account for the epistemic force delivered by the coherence of (a system of) beliefs. Contrary to widespread belief, however, this does not imply that Husserl subscribes to coherentism. In fact, Husserl does the opposite. For him, intuitions are not only a source of immediate justification, they are the ultimate source of justification. All we justifiably believe evidentially depends on our originally presentive intuitions. This makes him a foundationalist. As he acknowledges the fallibility of immediate, intuitive justification, he is a moderate foundationalist. More precisely, he subscribes to a version of moderate foundationalism we have specified in Sect. 1.2. as the most tenable version of foundationalism.