On the one hand, the basic distinction between insider and outsider is an obvious and useful one: it is undeniable that members of certain groups or strata have privileged access to knowledge, resources, and authority. On the other hand, in the study of religion, this basic empirical distinction has often been held to have deeper epistemological implications, i.e. asserting, as a necessary rather than contingent fact, that only insiders can have access to true knowledge of their faiths. Obviously, certain groups have their own subject matters and vocabularies that are relatively inaccessible to outsiders, e.g. quantum physicists or advaita vedanta gurus. If the matter of epistemic access were always relative in this manner, there would be no insider–outsider (I/O) problem, in any fundamental sense, in the study of religion; there would be no need for omnibus volumes reviewing the voluminous literature on the topic;Footnote 1 there would be only a well-known set of challenges, familiar to ethnographers and other outsiders.
However, some scholars of religion and religious practitioners argue that the case of the religious insider raises more fundamental barriers to access to knowledge for outsiders. On this view, the I/O distinction becomes the I/O problem, in that outsiders' epistemic access to the knowledge held of religious insiders is ruled out in absolute terms, not merely constrained or challenged by contingent social and cultural factors. Sometimes this is framed in terms of indoctrination or apprenticeship. For example, Darshan Singh argues that the ‘inner meaning of a religion unfolds only through participation by following the prescribed path and discipline’.Footnote 2
More commonly, it is framed in terms of the irreducibility of religious experience, which Wayne Proudfoot traces to Schleiermacher's insistence that ‘[i]mmediate intuition cannot be communicated’ (Proudfoot (Reference Proudfoot1985), 51). Rudolf Otto, for example, held that religious experience is rooted in a sense of the numinous, a
mental state [that] is perfectly sui generis and irreducible to any other … There is only one way to help another to an understanding of it. He must be guided and led on … through the ways of his own mind until he reaches the point at which ‘the numinous’ in him perforce begins to stir, to start into life and into consciousness. … In other words, our X cannot, strictly speaking, be taught, it can only be evoked, awakened in the mind … (Otto (Reference Otto and Harvey1958) [1917], 7)
In an externalized equivalent of this claim, Mircea Eliade held that
A sacred stone remains a stone; apparently (or, more precisely, from the profane point of view), nothing distinguishes it from all other stones. But for those to whom a stone reveals itself as sacred, its immediate reality is transmuted into a supernatural reality. In other words, for those who have a religious experience all nature is capable of revealing itself as cosmic sacrality. (Eliade (Reference Eliade and Trask1968), 12)
Insofar as the relevant experiences in such examples are taken as markers of religious belonging, only insiders can achieve this unteachable, supernatural knowledge.
Absolute views of the I/O distinction as a problem for outsider access to knowledge can also be framed in terms of modes of interpreting religious language. This can involve the claim that only insiders have certain qualities that are required for access to true interpretations. For example, the Hanbaliya school of Islamic thought insists that being a Muslim is a prerequisite for the quality of iman (acknowledgment or acceptance, distinct from belief) which in turn grounds true knowledge (Haj (Reference Haj2009), 47–50).
An emphasis on the symbolic nature of religious language – where only insiders have access to the means for this form of interpretation – produces relevant tensions between interpretative communities. For example, the Islamic concept of ‘inner discernment’ (fiqh al-batin) – along with a variety of interpretative principles – provided a basis for Muhammad Abduh's criticism of taqlid (‘unreasoned obedience to authoritative consensus’), informing his often innovative judgments as mufti of Egypt from 1899 to his death in 1905 (Haj (Reference Haj2009), 21–24, 78–82, 110, 117). Only a Muslim, and only a Muslim with a certain interiorized approach to reading the Qur'an and the Hadiths, could arrive at what he argued to be the correct view of these matters.
This strategy for claiming privileged access to insider knowledge is often premised upon the view that religious language involves an odd sort of symbolic or metaphorical reference to its objects, something that Nancy Frankenberry (Reference Frankenberry and Frankenberry2002) calls ‘the theology of symbolic forms’. Paul Tillich held, for example, that ‘symbolic language alone is able to express the ultimate. … The language of faith is the language of symbols’ (Tillich (Reference Tillich1957), 41, 45). Where this view of religious language is bolstered by claims that only members of a certain group are positioned to offer sound interpretations – e.g. by virtue of certain personal qualities or religious experiences – the I/O problem appears to emerge in a particularly vigorous form. For example, the Qur'an states that unbelievers ‘are the ones that Allah has cursed, so He deafened them and blinded their vision’; and ‘it is not eyes that are blinded, but blinded are the hearts which are within the breasts’.Footnote 3 This is more than W. C. Smith's famous claim that ‘No statement about Islamic faith is true that Muslims cannot accept’ (Smith (Reference Smith1981), 97). It is rather the claim that only Muslims have the type of vision necessary to see religious truths.
Given these views, it is important to distinguish three things: the I/O distinction (a relative one); questions raised by this distinction; and what is generally called the I/O problem (an absolute interpretation of the distinction). We argue here against the view that the I/O distinction can be an absolute one. That is, we argue against the view that something about the language or experience of religious insiders provides some insurmountable barrier to knowledge about the meanings of their language or practices.Footnote 4
Strong and weak insiderisms
It is useful to distinguish between weak and strong claims regarding insider privileges, what we will call weak and strong insiderism. Weak insiderism holds that ‘some groups have privileged access … to particular kinds of knowledge’; strong insiderism holds that ‘particular groups … have monopolistic access’ (Merton (Reference Merton1972), 11; original emphasis).
Weak insiderism poses methodological challenges to the study of religion, prompting scholars to ask hard questions about how to access knowledge when its distribution reflects social boundaries. Of course, framing this basic sociological fact in such general terms – insider versus outsider – is less useful than using more specific distinctions: e.g. sangha or priesthood versus laity; initiated versus uninitiated ritual participants; committed members, marginal members, versus non-members; etc.
Strong insiderism undermines the study of religion almost entirely, for example, implying that only shamans can study shamanism. In sociological terms, strong insiderism is suspect because it reduces the issue of identity to an overly simplistic contrast between belonging or not belonging. It fails to take account of degrees of belonging (e.g. complete insiders, partial insiders, occasional insiders, marginal insiders, complete outsiders, etc.), and it fails to take account of multiple and overlapping dimensions of identity. Given strong arguments against sociological formulations of strong insiderism, we might wonder why the I/O distinction remains an issue at all in the study of religion.
What we suggest is that beliefs about the nature of religious knowledge lead to a distinct form of strong insiderism. Whereas the sociological version holds that insiders have a monopoly on certain forms of knowledge by virtue of their social identity and/or position, the religious version holds that this monopoly arises from the distinct nature of the relation between the religious subject and the object of their knowledge. If we hold that there is a distinct mode of religious perception or that the object of religious knowledge is of a distinct sort, and if we hold that being ‘religious’ is a function of this mode of perception or this epistemic relation, then we have a strong version of insiderism based not directly on social belonging but on ‘religious’ knowledge itself. In essence, this is a transcendent argument for a sui generis view of religion: the sacred exists and must be known in a sacred way, therefore religion is a thing apart and must be studied as such. There is, of course, a circularity here: to be religious is to hold religious beliefs. We can further distinguish between universal and particular forms of the religious version of strong insiderism. For example, ‘religious people have a monopoly on religious knowledge because only they know the sacred’ versus ‘Roman Catholic Christians (or Lubovitcher Hassidic Jews or Nizari Isma'ili Muslims) have a monopoly on religious knowledge because only they know God's Revelation’. The argument for a religious version of strong insiderism need not be transcendent. For example, Brazilian neurosurgeon Raul Marino Jr. (2005) suggests that God causes brain states, including potential physiological differences that could explain atheists’ inability to acknowledge God's Revelation.Footnote 5 In the religious version of strong insiderism, it is the unusual nature of religious knowledge itself that grants insiders a monopoly on this knowledge.
This diagnosis of the nature of the I/O problem forces scholars of religion to address squarely some difficult philosophical issues. If we accept a weak version of insiderism, there is no ‘problem’, only an obvious, if difficult, set of methodological challenges. Problems arise with strong insiderism. The sociological version seems to fail on a number of grounds. This leaves us with the religious version. However, since it is explicitly based in claims regarding the nature of religious knowledge and meaning, it can only be defended, or critiqued, by recourse of epistemological and semantic theory. That is, the I/O problem in the study of religion is, at its heart, a philosophical one, and discussing its nature, or whether it even exists, requires us to use the tools of philosophy. In the remainder of the article, we explore some of the implications of this point.
Semantic theory and the insider/outsider problem
The use of the term ‘problem’ in the standard label is contentious. Jeppe Jensen denies there is a genuine I/O problem at all, declaring it a ‘manufactured’ pseudo-problem of ‘no methodological value’, merely a ‘politically expedient’ source of ‘exclusion’ (Jensen (Reference Jensen2011), 37–38). Less pessimistically, Thomas Ryba (Reference Ryba2007) thinks the problem to be either ‘pseudo’ or ‘profound’ depending on how it is construed (but he thinks that even its ‘profound’ formulations need not be worrisome to the religious scholar). What presuppositions are required for there genuinely to be a philosophical I/O problem with respect to theorizing about religion?
Let us start, somewhat arbitrarily, with Russell McCutcheon's characterization: the ‘problem’ is ‘to what degree, if any, … the motives and meanings of human behaviours and beliefs [are] accessible to the researcher who may not necessarily share these beliefs and who does not necessarily participate in these practices’ (McCutcheon (Reference McCutcheon1999), 2).
McCutcheon presents a dilemma with respect to the range of possible solutions: the scholars of religion either ‘have virtually unimpeded access to the intentions and meanings of the people, societies, or institutions they study’, or else are ‘cut off from ever being able to see past their own biases, contexts, and presuppositions’ (McCutcheon (Reference McCutcheon1999), 2). The ‘minds’ of those the researcher studies are either ‘open books’, or ‘enigmas’, or ‘blank screens’ on which the researcher can project her own beliefs and presuppositions.
Three things should be noted here. First, McCutcheon's formulation of the issues poses questions, not problems, for the study of religion. Second, it raises an embedded series of philosophical questions, most notably concerning the nature of mind, mental states, and intentionality. Third, it is, at base, a semantic formulation, making explicit reference to meanings and beliefs. In sum, if there is an I/O problem, then it will have a philosophical grounding, and the issue of meaning will be central. The question is whether, or to what extent, the agnostically inclined academic researcher can understand, in the sense of correctly interpret, the religious behaviours, beliefs, or experiences of religious practitioners, i.e. whether, or to what extent, the ‘outsider’ scholar of religion can understand or interpret the religious mental states of ‘insider’ religious adherents.
As such, the I/O question rests upon a core presupposition, namely that there is a fundamental dichotomy between first-person and third-person access to mental states. This presupposition generates a ‘problem’ only when we add a second one: that there is a difference in the authority or reliability of first- and third-person access.Footnote 6 In other words, first, there is a basic difference between knowing our own thoughts and those of others, and, second, we have more reliable access in one of these cases. This is reflected in the two predominant metatheories in religious studies – the ‘humanists’ who seek phenomenological understanding and the ‘scientists’ who seek causal explanation.Footnote 7 The I/O question becomes an I/O problem if there is reason to believe that there are fundamental impediments to genuine third-person access to the adherent's mental states. Any such impediment would have to take the form of preventing the scholar from correctly interpreting the mental states of the adherents, i.e. it would have to be such that the third-person access yields a semantic content distinct from that of first-person access. It is precisely this requirement which semantic holism challenges.
‘Semantic holism’, although used as a substantive, does not refer to a single, clearly articulated, theory.Footnote 8 All forms, however, will be unified against ‘atomistic’ semantic theories that postulate basic semantic units (words, sentences) whose meanings are in some sense self-contained or independent of the meanings of other semantic units. Holistic theories rather hold that the meaning of any given linguistic item is inseparably tied to the meanings of other linguistic items. These semantic relationships spread across an entire language, leading to the oft-repeated claim that, according to holism, the entire language constitutes the basic unit of meaning.Footnote 9 In a non-philosophical nutshell, if you think meaning lies in solitary words, you are an atomist; if you think it lies in a much broader contextual web of relations, you are a holist.
Davidson's version of holism, however, involves other, often more important, elements.Footnote 10 In many ways, he was most concerned with traditional philosophical problems – the reality of the external world and other minds, scepticism, moral action and weakness of the will, etc.Footnote 11 His philosophical importance was to argue, over a sustained period of time, that many of these ‘problems’ are either solvable or else turn out to be illusory by careful reflection on a most fundamental fact: that our ability to understand these to be problems, and our ways of thinking through solutions to them, are rational activities. Constraints on what it is to be rational, then, will impose constraints on the nature of the ‘problems’ and the range of potential solutions.
We note an ambiguity with respect to ‘rational’. Construed narrowly, rationality is an overtly normative concept – to be rational is to think well in some sense. Construed broadly, however, Davidson identifies rationality simply with the capacity for thought. For Davidson, thought itself is, at base, a linguistic activity. His view isn't so much that language is the medium through which thought and rationality is expressed, but rather that language is thought; that is, language without thought, and thought without language, is impossible: ‘Here we go in a circle: propositional thought requires language, language requires thought’ (Davidson (Reference Davidson and Caro1999), 42). Davidson uses the notion of ‘propositional attitude’ as the bridge between cognition and language; as he says, ‘to be a rational creature is just to have propositional attitudes, no matter how confused, contradictory, absurd, unjustified, or erroneous those attitudes may be’ (Davidson (Reference Davidson and Davidson1982), 95).
So, we find a tripartite identity implicit in Davidson: being rational=having propositional attitudes=being a language-user. Insights into the nature of language, then, provide insights into the mental states of the scholar's subjects. According to Davidson, language is, by its nature, communicable. The noises you make or the squiggles you put on a page are linguistic only to the extent that they are outward manifestations of ‘inner’ propositional attitudes, i.e. if they are such as to ‘communicate’ those attitudes. As alluded to above, communicability presupposes interpretability. At base, the evidence that the sounds emanating from you are indicative of your rationality is that they are interpretable – i.e. that hearers may, under specifiable conditions, identify the mental states that they purport to communicate. We can thus express a core thesis of Davidsonian holism as this: there is nothing more to meaning than what is required for interpretation.
Although some philosophers have concentrated on the ‘attitudinal’ side of propositional attitudes,Footnote 12 Davidson began his project with groundbreaking work in the ‘propositional’ side, with work on a ‘formal’ theory of meaning that he later integrated with a ‘material’ theory of interpretation.Footnote 13 While the nuances, complexities, and complications of this formal side of the Davidsonian project need not here concern us, there are two central points which emerge: (1) the ‘meaning-specifying theorems’ take the form of a relation between two languages – the ‘object’ language of the speaker and the ‘meta’ language of the interpreter;Footnote 14 and (2) there will necessarily be an indefinite number of alternative and non-equivalent proposals for meaning which satisfy these formal constraints. These ‘material’ theories must then be vetted for further adequacy.Footnote 15 These further constraints involve the methodological necessity of the Principle of Charity (POC) and the assumption of First Person Authority. Without going into the details, the POC requires that the interpreter assume that those she is interpreting are as rational as she. In other words, they have the same commitment to truth in their beliefs, consistency in their attitudes, ability to reason in accordance with basic logic, responsiveness to changes in the observable environment, etc. As a result, Davidson demystifies or dereifies meaning. In a nutshell, I interpret you by ‘translating’ your sentences into my own idiom. The ‘translation’ is correct when the propositional attitudes that are associated with the sentence in my idiom are (roughly) the same as those that are associated with the sentence in yours. This requirement presents a deep methodological puzzle – I correctly interpret your language by associating it with your beliefs (propositional attitudes), but the only access I have to what you believe is through interpretation of your language. Davidson's strategy for breaking into this ‘meaning-belief’ circle involves both the POC and the assumption of First Person Authority (more on this later).
Recall that the fundamental presupposition of any strong form of the I/O problem is that first- and third-person access to mental states respectively yields different semantic contents. Davidsonian semantic holism simply precludes this fundamental presupposition, and hence does not allow for a basis in which I/O perspectives can escalate to the level of I/O problems.
The argument for this derives from the fact that a strong I/O problem is tantamount to a thesis of the incommunicability of the adherent's language. Such a thesis would appear to render the adherent's language private in the sense which Wittgenstein's celebrated private language argument refutes. Wittgenstein describes a ‘private language’ as follows:
But could we also imagine a language in which a person could write down or give vocal expression to his inner experiences – his feelings, moods, and the rest – for his private use? … The individual words of this language are to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language. (Wittgenstein (Reference Wittgenstein and Anscombe1972), section 243)
There are two main elements in his exposition: the terms of the language refer exclusively to inner experiences; and such a language is untranslatable or uninterpretable by another language. The second aspect is strongly analogous (if not identical) to the core claim of strong insiderism, that insiders’ language cannot be understood by outsiders. It is thus tempting to take the widespread acceptance of the impossibility of such a private language as destructive of the coherence of strong insiderism.
However, this may be a bit hasty.Footnote 16 Taking the first aspect at face value suggests that a private language, if possible, would necessarily be uniquely understood by a single individual, whereas nothing in strong insiderism per se precludes the possibility of the insider's language being shared by others who are similarly ‘inside’. We can thus distinguish between a subjectivist and a social form of strong insiderism. On the subjectivist reading, only the insider herself can understand the semantic content of her own religious language, whereas on the social reading, only members of some group G can understand such semantic content. If some terms of religious language are taken to refer to some sort of (special) experience, subjective insiderism presupposes that such experience is only to be had by the individual, whereas social insiderism presupposes that such experience is common to members of G. Members of G can, therefore, converse quite intelligibly about their experiences with each other, but no non-G researcher can grasp the semantic significance of that language. It is, we think, an open question whether the core of Wittgenstein's reflections is limited to the subjectivist examples with which he introduces the problem. However, to move things along we will allow that while the subjectivist form of strong insiderism is imperilled by Wittgenstein's Private Language Argument the social form remains standing.Footnote 17
Davidson's holism, however, renders that formulation equally suspect. From a semantic point of view, the untranslatability or uninterpretability of the strong insiders’ language to those understood by outsiders is the crucial element. The plausibility of this rests on the coherence of the following claim: members of G speak a language L1 which is untranslatable into L2 spoken by non-members of G. Quite simply, Davidsonian holism renders this ‘claim’ incoherent, at least with respect to the prospects of an empirically grounded study of religion.
At base, this ‘claim’ presupposes a sharp divide between the identity- and the translatability-conditions for a language. That is, it requires that the insiders have a language that is untranslatable. With respect to an empirical study of such a religious tradition, the researcher must be able to recognize that the ‘subjects’ speak a language of whose semantic content she can have no inkling. But, Davidson's challenge is to ask on what basis – on what evidence – the researcher could grant the existence of such an untranslatable language.
After all, strong insiderism depends on the untranslatability of a language, not on the mere contingent fact that it has not actually been translated. Ancient Egyptian, as scripted in hieroglyphics, did not cease being a language between the fourth century and its decipherment via the Rosetta Stone.Footnote 18 In other words, what is required by strong insiderism is that there be no possible evidence on which expressions in the insider's ‘language’ can be understood as expressing the same thoughts (within the limits of indeterminacy) as those of the outsider's.
It is also crucial to understand the holist's conception of ‘translation’.Footnote 19 Translation is a relation between languages, but does not consist in a symbol-by-symbol or expression-by-expression correspondence.Footnote 20 Indeed, any ‘languages’ that were so mechanically translatable would in fact not be separate languages, but would be mere stylistic variants.Footnote 21 In other words, holists reject any syntactic account of translation, and insist on a semantic understanding: translation involves a process whereby the meaning (i.e. semantic content) of an expression in one language can be given by the meaning of an expression in another. Given the potential to misunderstand ‘translation’ as a syntactic transformation, as well as Quine's (Reference Quine1960) famous arguments for the indeterminacy of translation, Davidson prefers the term ‘interpretation’. For Davidson, language consists in its being a medium for expressing propositional attitudes, i.e. the thoughts, beliefs, desires, hopes, etc. of rational creatures. It is a species of intentional behaviour.
So, again, on what basis – by what evidence – can the ‘outside’ researcher conclude that her ‘inside’ subjects have a genuine but uninterpretable-from-the-outside language? A seemingly obvious answer is that they tell her so.Footnote 22 However, on pain of contradiction, the subject cannot be telling the researcher that she has an untranslatable language in that language, i.e. the proposal requires that the subject be possessed of two distinct languages – the accessible-to-all public (i.e. non-religious) one (‘P’) and the inaccessible-from-the-outside religious one (‘R’). While holism certainly has no problem with polyglottism, the proposal requires that the two languages be so semantically distinct as to challenge its plausibility. There are two relevant ways to conceive of such semantic differences.
The first involves a linguistic hierarchy, i.e. the claims are at different semantic levels. On the one hand, we have some insider object-level claim (of the religious language R), which ex hypothesi we cannot identify but at best only postulate its existence and pseudo-ostend it with a dummy name S. On the other hand, we have the meta-level claim (of the open language P) that says of S that it is untranslatable or uninterpretable. In other words, the interpretable meta-claim mentions the uninterpretable object-claim. Readers familiar with expressivism in metaethics (the view that the semantic content of moral language is limited to the truth-indifferent expression of emotions) will recognize the conditions in place for a Frege–Geach Problem (Geach (Reference Geach1965) ). Basically, S must continue to have the same semantic content when embedded in the meta-context, otherwise the meta-claim cannot assert that it is S which is uninterpretable (from the outside). But, as the meta-claim is publicly accessible, its constituent terms must have an accommodating semantic content, i.e. they must themselves be publicly accessible. Therefore S, a constituent of the meta-claim, cannot have an inaccessible content. Conversely, if S does have inaccessible content, then so must the meta-claim. The proposed solution – that the subjects tell the researcher that they have an uninterpretable-from-the-outside language, and that that claim constitutes the sole evidence of such a language – is refuted.Footnote 23
The second semantic difference is that the two languages would have to be semantically isolated from each other, in the sense that they could not be mutually interpretable. That is, translation would fail as a symmetrical relation. While it is an open question whether the closed religious language R can be adequately used to interpret open public language P, it is clear that P could not be used to interpret R. Suppose it could; then R would not, contrary to the hypothesis of strong insiderism, be uninterpretable from the outside – outsiders could easily interpret it using the resources of the openly accessible P.
Besides raising an analogous problem to that involving linguistic hierarchy (i.e. if P cannot interpret R, it is difficult to understand how, nonetheless, P can be used to express important truths about R), more holistically focused challenges to this second view emerge. Holism is predicated on the idea that the meaning of any linguistic expression depends, at least in part, on its relation to the meanings of others. In its traditional expression, meaning forms a ‘web’ spreading outward from any particular node in multiple and complex ways. The logical result is that the meaning of any node cannot be held a priori to be independent from the meaning of any other (though, of course, nodes will lean more heavily on those closest to them). The semantic isolation required between nodes of P and R, however, threatens this basic holistic insight; insiders must be diagnosed with linguistic schizophrenia of a rather implausible sort, at least of a sort which would make the proposed solution – that insiders tell the researcher than they have an uninterpretable language – performatively impossible. Taking Davidson's unity of meaning and thought seriously, the linguistic schizophrenia would also be tantamount to a doxastic one as well, i.e. on such a view no religious belief could have any non-religious content.Footnote 24 Such bifurcations seem antithetical to the whole spirit of semantic holism. To the extent that one feels the holistic impulse at all, one should seriously question the reliability of insider's claims to the existence of an uninterpretable language.
What other basis might there be that the insider possesses an uninterpretable-from-the-outside language? Given the implausibility of semantic bifurcation that emerged from discussion of the previous proposal, and given Davidson's insistence that any form of interpretation be constrained by what is required for radical interpretation, the question seems to be a demand for evidence that the ‘subject’ is a language-user at all. In other words, the case for strong insiderism requires the existence of insiders whose totality of language use, beliefs, and/or intentional actions are uninterpretable in principle by any outsider, and the problem is to provide an evidential basis for the existence of such creatures.Footnote 25 In other words, what is required is the kind of evidence which, in the complete absence of linguistic interpretation, a researcher would need in order to conclude that the creature she is observing is a language-user. Two prima facie candidates spring to mind: (i) the subject is a human being, and human beings are, by nature, language-users, and (ii) the subject behaves in ways observationally indistinguishable from the behaviour of uncontested language users.
Consider the first candidate, that the researcher recognizes the humanity of the creature in front of her and accepts that humans are, by nature, linguistic. The clear equivocation is on the expression ‘recognizes the humanity’. One recognizes the humanity of another in a biological sense when one recognizes membership in the genus/species homo sapiens. Ignoring the etymological implications of the term ‘sapiens’, this is simply not the same as recognizing the rationality (=intentionality=possession of propositional attitudes=language-use) of the other. This is not to say that recognition of homo sapiens is irrelevant to recognition of rationality, but only that it radically under-evidences it. Conversely, we may come to a conclusion about rationality even in the absence of humanity (e.g. great apes, angels, corporations).
Consider the second candidate, that the researcher recognizes complex behaviour from the subject indicative of linguistic ability. Obvious examples include vocalizations with recognizable and systematically categorizable patterns. Certainly, such behaviours are evidence of linguistic-ability; any one denying this is on very thin ice. Nonetheless, it is crucial to distinguish this evidence of language from the language itself; the language does not consist in the marks on a page or the sounds in the air; it involves the propositional attitudes which those writings and vocalizations express. The marks and sounds are arbitrary and only conventionally linked to what they express.Footnote 26 In other words, such complex behaviour similarly under-evidences rationality and linguistic ability.Footnote 27
It might be suggested that our argument has an unreasonable standard of evidence, and that a Davidsonian critique of strong insiderism grasps at implausible straws. In response, we would point out that Davidson himself, as much as anyone, trumpets the importance of the recognition of humanity and complex vocalizations: they play ineliminable roles in his account of radical interpretation, especially in the application of the all-important POC. It is true that Davidson is committed to the idea that the outsider can have no reason to believe the creature in front of her is rational other than on the basis of being able to interpret its behaviour as rational. However, this most-Davidsonian of theses can be mis-emphasized. It is not that we encounter biological humans with the assumption that they are non-rational until proven otherwise. Quite the contrary: we encounter them with a cluster of assumptions united under POC, e.g. that they are similar to ourselves, that they are rational, that they speak a language, that they have mental lives, that the commonalities between their mental lives and our own dwarfs the differences between them, etc. In other words, POC is the prevailing presupposition that is only jettisoned in the face of prolonged and persistent failures of communication. The crucial point is this: strong insiderism guarantees that POC will fail to provide a basis for interpretation, and it is for this reason that it guarantees the incoherence of the outsider simultaneously viewing the insider as speaking a language at all, and hence guarantees the incoherence of the outsider's being able to view the insider as speaking an uninterpretable (by her) language.
We submit that the sort of complex behaviour in observable circumstances which the outsider would display, in the expression of her own beliefs and desires, constitutes the sole evidence on which she can conclude that another is speaking a language. In other words, the only such evidence is that the insider's total behaviour is interpretable by the outsider. This is enough to show an irreconcilable tension in the idea that there may be something identifiable as an untranslatable language. As the very idea of strong insiderism presupposes the existence of untranslatable languages, strong insiderism is precluded on Davidsonian grounds.
There is an important consequence or corollary here to Davidson's position. Recall that, on his view, there is nothing more to meaning than what is required for (radical) interpretation, where interpretation is thought of as a ‘translation’ from one language to another. Translation and interpretability are symmetrical relations. If first- and third-person access yielded different contents, then there would be no interpretation from the adherent's language to the scholars, nor from the scholar's back to the adherent's. As such, there would be strong grounds for denying the meaningfulness of either. In other words, Davidson's holism is committed to denying any semantic priority or authority on the basis of the form of access to the mental states. In still other words, for Davidson, semantic content is more basic than semantic access, and hence the nature of the access cannot alter the nature of the content.Footnote 28
In many ways, we anticipate this to be the most controversial and complicated part of our argument, but in the space remaining we can only consider one obvious objection, namely that Davidson's form of radical interpretation itself relies fundamentally on what he calls the assumption of First Person Authority. There is a good deal of very interesting and important critique of Davidson on this point: for example, Matthew Day (Reference Day2004) and Terry Godlove (Reference Godlove and McCutcheon1999) both call that assumption into question at least with respect to religious belief. However, the very label that Davidson uses gives rise to the confusion. Throughout this article the idea of ‘semantic authority’ has been the idea of whose access, and hence whose understanding, should be taken as authoritative with respect to interpretation (e.g. if the adherent understands the significance of her ritualistic actions in a manner different from that offered by the scholar, then whose trumps whose?). However, this is not the sense of ‘authority’ required by Davidson. All he means by it is that when we speak we know what we mean in a way that we do not when others speak. Specifically, we must ‘interpret’ or ‘translate’ the utterances of others, but not our own, by implicit reference to constructible meaning-theories.
In terms of testing a proposed meaning-theory – i.e. attempting to interpret – the assumption of first-person authority allows the interpreter to break into the ‘meaning-belief circle’ in her own case, and the assumption of charity allows her to bootstrap to the third-person case. If this is the basic form that interpretation takes, informing the very concept of semantic meaning, then we see that both a first-person (=insider) and a third-person (=outsider) perspective are required for the very possibility of meaningfulness. The I/O distinction, then, is ubiquitous in each and every act of interpretation.
Conclusion
In sum, when we understand meaning in terms of a broad web of interpretation, not something that attaches to individual bits of language, there simply is no I/O problem, in the absolute sense in which this is often discussed in the academic study of religion. Davidson's view of meaning rules out an I/O problem because insiders and outsiders both mean what they mean in public ways that are interpretable in basically the same way. The only reason to treat insiders as immune to interpretation would be to treat them as completely non-rational. In the light of a holistic semantics, strong insiderism, in both its sociological and religious versions, fails.
We are left with the point with which we began: the basic empirical distinction between insider and outsider is an obvious and useful one. It is a basic fact of life that members of a given group tend to have privileged access, if only through longer acquaintance, with certain sorts of discourse, beliefs, and practices. However, holistic views of meaning rule out claims that religious insiders have radically privileged access to certain forms of knowledge and that this access is not at all available to outsiders. This is not merely a minor point regarding a somewhat idiosyncratic debate in the study of religion. Semantic holism emphasizes the far-reaching scope of this distinction: the insider–outsider distinction applies to all acts of interpretation in all possible contexts; weak insiderism is an inherent feature of all communication; and strong insiderism is similarly ruled out in all possible contexts.Footnote 29
