Phenomenology and the Cognitive Sciences

, Volume 12, Issue 2, pp 327–336

Linguistic competence and expertise



DOI: 10.1007/s11097-011-9211-5

Cite this article as:
Addis, M. Phenom Cogn Sci (2013) 12: 327. doi:10.1007/s11097-011-9211-5


Questions about the relationship between linguistic competence and expertise will be examined in the paper. Harry Collins and others distinguish between ubiquitous and esoteric expertise. Collins places considerable weight on the argument that ordinary linguistic competence and related phenomena exhibit a high degree of expertise. His position and ones which share close affinities are methodologically problematic. These difficulties matter because there is continued and systematic disagreement over appropriate methodologies for the empirical study of expertise. Against Collins, it will be argued that the term ‘expertise’ should be reserved for expertise (esoteric experts) and exclude everyday performance (ubiquitous experts). Wittgensteinian ideas will be employed to maintain that it is mistaken and misleading to derive substantive conclusions about the epistemology of expertise from ordinary linguistic competence and vice versa. Significant attention will be devoted to the notion of following a rule with particular reference to the intelligibility of tacit rule following. A satisfactory theoretical approach to expertise should not involve making important claims about ordinary linguistic competence.


ExpertiseCollinsLinguistic competenceRulesWittgenstein


There is continued and systematic disagreement over appropriate methodologies for the empirical study of expertise. Developing a consensus over suitable approaches is important for progressing interdisciplinary research. Harry Collins’s extensive writings offer a distinct way of investigating questions about expertise. A study of his work indicates that his philosophical commitments have direct implications for the methodologies he favours, but for reasons of space, the links between the two will not be discussed. Investigation of selected aspects of Collins’s programme indicates that there are significant philosophical problems which generate methodological difficulties. An essential part of his research programme “Studies of Expertise and Experience” (SEE) is that there is both esoteric and ubiquitous expertise. In support of this view, Collins places considerable weight on the argument that ordinary linguistic competence and related phenomena exhibit a high degree of expertise. As he puts it (Collins 2011):

While traditional analyses take the word ‘expert’ to refer only to rare, high-level specialists, SEE considers that ordinary language-speaking, literacy, and the like exhibit a high degree of expertise even though everyone has them– they are ubiquitous. That is, perhaps, one of the most radical contributions of SEE to the analysis of expertise as indicated by initial strong opposition to the idea of ‘ubiquitous expertise’ from philosophers and psychologists.

Against Collins, it will be argued that the term ‘expertise’ should be reserved for expertise (esoteric experts) and exclude everyday performance (ubiquitous experts). Expertise is not ordinary competence or performance and extending the former so that it encompasses the latter is neither warranted nor enlightening. It is worth observing that the definition which SEE has of expertise is almost exclusively adopted by those working in this particular research paradigm and is not shared by the vast majority of other researchers in the field (however, for a similar view to SEE, see Cianciolo et al. (2006), p.14), which allows for the idea of expertise being demonstrated by the majority of individuals when they succeed in their everyday lives). An extension of the definition of expertise in this way, especially against a very broad consensus opposed to it, requires strong argumentation coupled with explicit recognition that the meaning of the word ‘expertise’ is being altered. In this respect, Wittgenstein’s writings about meaning change are germane. He thought that altering the meaning of a word is acceptable, but a linguistic community must be aware that certain words are being used in new ways in order to avoid philosophical confusion (including mistaking a different way of talking for a scientific discovery). Wittgenstein remarked (Wittgenstein 1972, p.23):

Thus, by the expression ‘unconscious toothache’ you may either be mislead into thinking a stupendous discovery has been made, a discovery which in a sense altogether bewilders our understanding; or else you may be extremely puzzled by the expression (the puzzlement of philosophy)…

The misleading temptations which arise from ‘unconscious toothache’ clearly directly parallel those raised by ‘ubiquitous expertise’. It might be thought that some of the apparent plausibility of Collins’s position trades on an equivocation with the usual meaning of ‘expertise’. He needs to do far more to allay these concerns if his extension of ‘expertise’ is to be at all persuasive.

Esoteric and ubiquitous expertise

To fully appreciate Collins’s reasoning for his view, it is useful to examine a number of general arguments he gives at the start of Rethinking Expertise. He writes (Collins 2007, pp. 2f.):

It starts from the view that expertise is the real and substantive possession of groups of experts and that individuals acquire real and substantive expertise through their membership of those groups. … In the case of relational theories, on the other hand, all the work is done by the attributors. Under our treatment, then, individuals may or may not possess expertise independently of whether others think they possess expertise.

He holds that relational theories of expertise deem expertise to be a matter of the relationship which experts have with others. The attribution of expertise by the (frequently retrospective) assignment of a label is an instance of a relational theory. Collins’s position is unsatisfactory because it confuses the logical possibility of ascribing expertise independently of what others think with the epistemological plausibility of such an attribution. In many cases, it is hard to see how to sensibly ascribe expertise independently of what people think. For example, if it is wondered whether a consultant in cardiac surgery is an expert in normal circumstances a key part of this ascription will be peer judgement. Another instance is expertise connected to political activism and the fact that part of what enables activists to have a significant role in the political process is the recognition by other people of their expertise. Epstein (2000) is a representative discussion of such considerations. Collins presents a very extreme form of the relational theory, and thus it is possible to see what might be thought objectionable about it but taking adopting this line of argument does nothing to demonstrate that a more restrained version of the relational theory is seriously flawed. It is not all obvious why his position that the attribution of expertise is independent of what others think is more plausible than a moderate rendering of the relational theory. A version of this sort allows for the possibility of ascribing expertise in a manner that is not necessarily and exclusively tied to whether others think that individuals have expertise.

Collins proceeds to illustrate his claim that relational theories are mistaken by an example. He states (Collins 2007, p.3):

To give a simple example, in France everyone can speak French, `even the little children,’ and it is not thought of as an expertise. On the other hand, in Britain a person who is fluent in French is thought of as an expert and can, for example, command a salary as a translator or teacher. … In a purely relational theory the expertise involved in speaking French and English is no more nor less than that attributed to speakers of the languages in their respective countries. In a realist/substantive analysis, on the other hand, the degree of expertise in speaking a language remains the same in whichever country the language is spoken.

This instance exposes some of the philosophical assumptions Collins relies upon and thereby makes his objections to relational theories less persuasive. His position is troubling on a number of significant grounds. Firstly, what is at stake in the situation Collins describes is not a matter of epistemology on any normal understanding of it. Furthermore, the fashion in which he handles it is a good illustration of a philosophical problem being created by a resolute refusal to accept an ordinary comprehension of the situation as being adequate. Being clear about whether an individual is an expert on French or English is a matter of clarifying the context if this is not already obvious and in many cases the context will be readily apparent for reasons to do with semantics and/or pragmatics (such as Grice’s maxims governing conversational logic). The unsatisfactory nature of what Collins claims can be further indicated by considering a few sentences which ascribe expertise. It would be odd to say ‘Pierre is an expert French speaker’ if he is French and not an expert on some aspect of the French language. That is to say, it would not be usual to attribute expertise to a native speaker who speaks their language in a way that would normally be regarded as competent. It should be recognised that instances of this kind do not preclude the sensible ascription of expertise in more unusual circumstances, but what makes these circumstances somewhat different from the norm can be recognised. In a related way, the sentence ‘Steve really is an expert walker’ would be misleading if it were used to mean that he walks normally. However, it would be appropriate to say this if he had developed some expertise to overcome a medical condition which almost invariably prevented walking or in a case where the selection of an Olympic walking team was being discussed and his expertise was being balanced against his tendency to suffer from nerves in big competitions. What this discussion shows is that resolution of questions about the attribution of expertise in the example Collins gives requires clarification of meaning and not the development of epistemological theory. Secondly, Collins’s example has the potential to be profoundly misleading because it conflates first and second language learning. In recent years, there has been a huge amount of research on the nature of second language learning (in considerable part due to the increasing importance of English as a second language). In very broad terms, this body of work indicates that there are fundamental differences between first and second language learning. It follows from this that Collins’s example fails to demonstrate anything of significance about the acquisition of linguistic expertise and thus about the possession of expertise more generally. It is evident that neither the argument nor the example presented by Collins has shown that a moderate version of the relational theory should be rejected in favour of his position.

Collins relates the ubiquitous nature of expertise to the acquisition of natural language. He remarks (Collins 2007, p. 3):

Equating the ubiquity of an expertise with the absence of an expertise has been responsible for some serious mistakes. For example, before attempts were made to make natural language-handling computers the deep difficulties of language speaking were not understood. During that period a mistake was being made in treating native language speaking as easy, a mistake that revealed itself soon after attempts were made to construct machines to do it. Relational or attributional theories have no way of discussing whether language-handling computers really are capable of performing as advertised or why they might not be.

His argument is problematic for several reasons. To begin with, the assertion that equating the ubiquity of an expertise with the absence of an expertise is the cause of significant error is unjustified. This becomes apparent if the notion of ubiquitous expertise is rendered by that of ordinary competence or performance in the way that expertise researchers who do not share the SEE paradigm favour. If this substitution is performed the assertion becomes equating an ordinary competence or performance with the absence of an expertise has been responsible for some serious mistakes. There is nothing wrong with stating this as it is so general there must be instances which satisfy it. However, its generality has the drawback that it is not all clear anything of significant interest is being claimed. It is evident that a suitable response to this line of argument cannot take the form of maintaining that the notion of ubiquitous expertise cannot be expressed in any other terms because such a reply would merely amount to a restatement of the favoured position. Furthermore, as with the example about expertise in French and English, this case of computers handling natural language does not demonstrate the point Collins thinks it does. There is a significant research in the philosophy of artificial intelligence which is concerned with the difficulties generated by attempts to enable computers to handle natural language with notable issues being the Chinese room and frame problems. Whatever view is taken about the solutions of the Chinese room (Searle 1980) and frame problems (Dennett 1984) it is evident that the questions they pose are entirely distinct from those about the acquisition of natural language in humans and to conflate the two is misleading. It follows from this that Collins’s example has no bearing on the assessment of the merits of relational theories of expertise.

A difficulty with Collins’s view of expertise is that he is committed to a wholly sociological definition of an expert. His approach could be objected to the grounds that such definitions of expertise rely on the ability to make and sustain very refined sociological distinctions. Problems about the identification and boundaries of relevant linguistic communities for instances of expertise reinforce this criticism as Collins maintains that the recognition of these depends upon linguistic socialisation. It might be questioned whether the concept of linguistic socialisation, that is to say immersion in the relevant linguistic community alone (Collins and Evans 2007), is practically or even theoretically acceptable. The difficulty with the notion of linguistic socialisation is that following Wittgenstein’s ideas on language games (particularly Wittgenstein 1972, 1984) how is it to be decided what the relevant linguistic community is and how its boundaries are to be demarcated. Wittgenstein compared language with games, and the point of the analogy was to draw attention to various similarities between language and games. He extended the analogy to encompass language as a whole, and the term ‘language game’ stemmed from this. The concept of language games was designed to capture and express Wittgenstein’s ideas about the flexibility and variety of language and the role of the linguistic community (Parallel difficulties to those with the idea of linguistic socialisation arise in the case of so called religious language games of the kind propounded by D.Z. Phillips. See Addis (2001) for a treatment of this). If the boundaries cannot be appropriately demarcated, it is not clear how much explanatory force the mention of the relevant linguistic community has. What deserves further clarification is whether a linguistic community is taken to be synonymous with a cultural community and/or a community of practice. If it is the latter case, immersion would also seem to consist of involvement in cultural and/or communal practices. This is not necessarily problematic, but if it is so, it should be clearly indicated as the theoretical relationships between language and culture merit further development. Such considerations support the view that a definition (or preferably a characterisation) of expertise which allowed more room for psychology would be beneficial as this would reduce the weight placed upon sociological discrimination. A further problem for Collins is that his position that the definition of expertise is entirely sociological does not sit especially easily with his claim that the ascription of expertise is independent of what others think. If it is possible for individuals to have expertise independently of what people think then evidence for possession in such cases cannot be of (a primarily) sociological sort and so must be of a psychological kind. Difficulty arises because Collins clearly heavily favours an approach to expertise that is characterised by reference to behaviour and/or cashed out in behavioural terms. He is thereby committed to the very strong methodological claim that no psychological evidence can independently count towards the possession of expertise. It is evident that Collins must admit the possibility of psychological evidence in order to explain certain cases of having expertise at the same time as maintaining the view that no such evidence can count independently.

Problems about the role that psychological evidence should have in the attribution of expertise stem from wider issues about the extent to which expertise can be categorised in purely behavioural terms. As such, these questions belong to a traditional major strand of philosophical enquiry about the relationship between the inner and the outer taken in its widest sense. Too much emphasis on either the inner or the outer is philosophically problematic. Collins’s tendency to espouse the view that expertise is characterised by reference to behaviour and/or cashed out in behavioural terms leaves him in the position of being too insistent on the primacy of the outer. Furthermore his perspective does not acknowledge any innate talent or intelligence element to expertise even though such capacities affect the development of expertise. Collins’s reading of Wittgenstein tends to favour the strand of interpretation proposed by David Bloor (2002) and Peter Winch (1990) that strongly emphasizes the contribution which the former could make to the sociology of knowledge at the expense of underplaying the useful contribution which suitably understood psychology could provide. In contrast to this, it will be argued that there are good Wittgensteinian reasons (Hacker 1993) for claiming that neither the inner nor the outer takes priority, but what matters is explaining the relationship between them. What is problematic is not psychological investigations in themselves but the philosophical attitudes taken to such inquiries. Such a stance allows a role for investigations in psychology which seek to understand expertise through the study of individual capacities and differences, including the ability to learn. For example, it would seem reasonable to think that fully adequate accounts of the acquisition of expertise require some kind of explanation about the ability to learn and those which lack this do not seem to be wholly satisfactory. According more scope to psychology (particularly learning theories) would be an improvement on Collins’s approach to expertise.

Types of tacit knowledge

Collins provides a detailed classification of the types of tacit knowledge starting with the broad categories of somatic tacit knowledge, contingent tacit knowledge and collective tacit knowledge (Collins 2007). Somatic tacit knowledge is divided into somatic-limit tacit knowledge (Collins 2007), somatic-affordance tacit knowledge (Collins 2010) and non-declarative knowledge. Contingent tacit knowledge is further identified as unrecognised knowledge (Collins 2001a), unprompted knowledge, uncovered knowledge and unrecorded knowledge. Collective tacit knowledge (Collins 2007, 2010) is where what is tacit is due to its location in the social collectivity. Such knowledge allows individuals to carry out actions which require an understanding of the social context for their correct performance, and it includes rule formation and alteration (Collins 2001b). A general concern which this classification raises is that given tacit knowledge is, to a large degree, context dependent, it might be wondered how easy it is in practice to reliably discriminate between contexts so that instances can be clearly ascribed to a particular category. To put the point another way, the difficulty might be that if there is to be reliable ascription of cases, such a level of detail is required in the specification that it undermines the rationale for having the larger categories in the first place. Emphasis should be placed upon the development of suitable methods of empirical investigation to attempt to achieve greater certainty about what belongs to a particular category. There are potential dangers of reification in setting up such categories since they begin as conceptual conveniences and gradually become independently existing entities. Another significant cause for concern is that it is not at all obvious these categories belong to either of Hume’s relations of ideas or matters of fact thereby running the risk of their becoming a dubious species of metaphysics. Of course, it is not necessary that these categories should be augmented in a metaphysical direction, but if this is to be avoided, there should be a clear sense of what would constitute appropriate investigative procedures. Defending metaphysical or conceptual boundaries matters less than developing appropriate methodological principles particularly with respect to what is regarded as legitimate explanation.

The concept of collective tacit knowledge is a highly contentious one not least because it is not amenable to codification, and there is no systematised discussion about what should be included in this category. However, it might be suggested that questions about collective tacit knowledge should be raised at a fundamental level as many philosophers would deny its existence. In the majority of traditional epistemological models which offer analyses of the claim that person S knows proposition P, it is very difficult to find a theoretical space where such knowledge could be located. However, the problems of locating this theoretical space do not in themselves show that there is no such thing as collective tacit knowledge and indeed might point to taking epistemology in a radically different direction such as that offered in Wittgenstein’s On Certainty. Collins states that the reasons for the tacitness of collective tacit knowledge are ‘more ontological rather than biological’ (Collins 2007, p. 260). His view about the grounds for the tacitness of such knowledge is troubling because the reference to ontology seems to be leading in the direction of metaphysics and away from empirical inquiry. If the most important feature of collective tacit knowledge is how individuals come to possess it, this invocation of metaphysics does not appear to be particularly helpful and major unresolved questions about its nature remain. An underlying difficulty for the notion of collective tacit knowledge is simply that of explaining how something can be both tacit and be knowledge in terms of stating a clear contrast with what it would be for such knowledge to be fully explicit. For example, problems are posed by schemas of practical inference (Carr 1981) which take the form of:

It is A’s purpose to φ

A understands that ψ-ing is a way to φ

A ψs

Here, it is not clear what would count as tacit and as explicit with such difficulties being seriously augmented in instances where such schemas formed part of what was deemed to be collective tacit knowledge. These criticisms of Collins’s conception of the nature of collective tacit knowledge should not be interpreted as supporting the claim that group expertise does not exist. Indeed, it will be argued that given present levels of understanding of collective tacit knowledge, it is preferable to employ the concept of group expertise instead wherever it is possible to do so. At this point, it might be objected that the notion of group expertise could be formulated in terms of collective tacit knowledge. However, a reasonable response to this is simply to maintain that group expertise is a useful notion and that no further theoretical characterisation of its nature is desirable until more is known about precisely what it is. What is being argued for here is that it would be advantageous to move away from potentially irresolvable theoretical disputes about the nature of collective tacit knowledge and towards more operationalisable kinds of questions. For example, the effects of social identity and knowledge quality on knowledge transfer across groups could be investigated (Kane et al. 2005). It is possible if there was a great deal more known about the differences between individual and group knowledge, it would be easier to characterise what (if anything at all) collective tacit knowledge is.

Rule following

As part of his account of the nature of collective tacit knowledge, Collins makes significant use of Wittgenstein’s work (especially interpretation which Winch offers) and his own views about what computers are capable of to claim that rule formation and alteration is an ability which only members of a collectivity have (Collins 2001b). It might be questioned whether invoking this ability actually contributes anything in terms of understanding given a commitment to rule following in the Wittgenstein sense (Baker and Hacker 1994, pp. 154–181). Although there is scholarly dispute about whether an individual isolated from a community can follow a rule, the baseline interpretation is that rule following is a communal activity. It would therefore seem to be unnecessary to regard the ability to form and alter rules being significantly distinct from Wittgensteinian rule following in general. It might be objected that what Wittgenstein does or does not say about rule following is not particularly relevant. That is to say, some people agree with Wittgenstein and whilst others reject his views. However, this rebuttal is not really open to Collins because Wittgenstein has been invoked in a substantive way.

In considering what is involved in following a rule, Collins rightly focuses on the idea of establishing what is correct. However, significant difficulty is generated by his omission of Wittgenstein’s idea that following a rule involves understanding that one is following it and without this one is merely acting in accord with a rule. Wittgenstein insisted upon the distinction between actually according with a rule and merely believing that one is. He claimed that being in accord with a rule is not determined by whether an individual thinks he is in accord but by actions in practice which accord with it. For Wittgenstein, to follow a rule an individual must intend to follow it and be able to invoke it in account of his practice if required. For instance, a rule might be cited as an explanation for particular actions. Collins’s disregard of the difference between rule following and according with a rule is not just important in terms of appropriate Wittgenstein exegesis. His omission of the difference between the two is likely to lead to confusion between questions of judgement and those of rule following. Wittgenstein (1984, section 242) claimed that a linguistic community must agree in judgements (or at minimum in a core of these judgements). Such agreement is accord over what counts as the correct applications of a rule. The distinction between matters of judgement and rule following is significant because it is not always obvious what counts as judgements of sameness in rule following (see below). A further problematic consequence of Collins’s lack of differentiation between rule following and accord is that his notion of tacit rule following can be argued to be incoherent on a Wittgensteinian basis. This is because following a rule involves understanding that one is following a rule, and without this, one is acting in accord with a rule. It is important to appreciate that this is a conceptual and not an empirical point (Baker and Hacker 1994, pp. 154–181). Acting in accord with a tacit rule is possible, but following it is not. Invocation of the notion of applying a rule does not in itself really negate the importance of this point because it remains unclear whether application is following a rule or merely acting in accord with it. These criticisms of the concept of tacit rule following are not intended to impugn empirical data about how rules are followed in particular cases and what is required is that the data should be theorised in a different way.

A question which arises with rule following is what counts as a judgement of being the same in particular circumstances (or alternatively put what constitutes agreement in judgement in a given instance). Collins shows awareness of the problem as he claims that the idea of seeing something as the same needs the definition of an ‘area of tolerance or indifference to which the actor does not attend’ (Collins and Kusch 1998, p. 47). Given the significance which he ascribes to Wittgenstein, it might have been useful to supplement the discussion of judgements of similarity and difference with what is usually referred to as the issue of rule following in hard cases. There have been substantial attempts (with a good proportion of this stemming from jurisprudence (such as Bix 1995) as the application of rules of cases is a significant issue) to extend Wittgenstein’s treatment of rule following to cases where it is difficult to determine whether a rule is being followed or not. Without a substantial attempt to frame further questions about the nature of judgements in hard cases, it seems as though Collins is stating the problem rather than resolving it. In this situation, his view that rule formation and alteration is an ability which only members of a collectivity have remains in need of additional explanation since one of the questions which needs to be addressed is how rule formation and alteration functions in hard cases.

Collins’s attitude to philosophical methodology is well summarised when he writes (Collins 2011):

Anything that involves `loyalty’ to a philosophical cannon belongs, at best, in the lobes. Loyalty is best treated with suspicion. Misplaced loyalty to methods and perspectives has damaged sociology and, I suspect, it has damaged psychology. What matters is what works.

However, in contrast to his view, it has been argued that questions of methodology are not so easily put aside and that a number of significant aspects of Collins’s programme encounter significant philosophical difficulties which lead to methodological problems. Progressing interdisciplinary expertise work involves developing a suitable understanding of why theoretical approaches to expertise which involve making important claims about ordinary linguistic competence are fundamentally unsatisfactory and why research programmes founded on that basis are likely to fail to fulfil their promise.

Copyright information

© Springer Science+Business Media B.V. 2011