Encyclopedia of Educational Philosophy and Theory

2017 Edition
| Editors: Michael A. Peters

Wittgenstein and the Path of Learning

Reference work entry
DOI: https://doi.org/10.1007/978-981-287-588-4_404


The teacher/pupil relationship is ubiquitous in Wittgenstein’s writings, but quite how he understood it and what goes on in the transactions that enable learning is not clear. The key to understanding Wittgenstein on learning is to understand the characteristics and skills that his remarks require of the pupil qua individual. Most commentators read Wittgenstein as endorsing a social model of the teacher/pupil relationship: the pupil learns by being initiated into a social practice by the teacher (Stickney 2008; Williams 1994, 2002, 2011; Smeyers and Burbules 2006; Bakhurst 2011).

The first section sketches four reasons for challenging the social reading of Wittgenstein. The second section outlines the shape of an individualistic account of the path to learning.

Challenging the Social Reading of Wittgenstein

  1. 1.

    Training. The English word “training” is elastic in meaning. It covers all sorts of learning schedules from simple Stimulus–response (S-R) conditioning to initiation into complex activities that require sophisticated thoughtfulness, from learning nuanced craft skills to intellectual pursuits such as chess. Wittgenstein’s word for “training” is Arbrichtung and this has none of the elasticity of the English word. It applies only to crude S-R conditioning. It’s a word suitable for conditioning “dumb brutes” – for whipping horses. It is a not a word that is suitable in German for human training (Huemer 2006). At the very least, this means that we must treat Wittgenstein’s talk of training with extreme caution and would do well to assume that it means only simple conditioning. If learning involves training in Wittgenstein’s sense, it must include a good deal more too. Learning cannot consist simply in training; it must involve training plus something else (Stickney 2008, Luntley 2008, 2012).

  2. 2.
    Wittgenstein nowhere endorses a social account of practices; he nowhere says that a practice is constituted by being a shared activity. As David Pears pointed out long ago, in the one passage where his interlocutor directly raises the question, Wittgenstein ducks the issue, for he focuses on the need for repeatability over time, not across persons. In PI 199, his interlocutor says,

    Is what we call “following a rule” something that it would be possible for only one person, only once in a lifetime to do?

    to which Wittgenstein replies:

    It is not possible that there should have been only one occasion on which only one person followed a rule. It is not possible that there should have been only one occasion on which a report was made, an order given or understood, and so on.

  3. 3.

    Wittgenstein’s speaks of practices, habits, and customs as part of our “natural history” (PI 25, 415), but he nowhere says of these things that they are normative, nor does he speak of norms. These terms are not part of Wittgenstein’s vocabulary. He speaks of rules – Regln – but this word is, of course, from the same root as Regularitie. What Wittgenstein is interested in are regularities (PI 208). The idea that the regularities or word use are normative or involve “grasp of norms” is a fiction of the secondary literature that has dominated much writing on Wittgenstein since Kripke’s seminal work on rule-following. Furthermore, the idea that word use – linguistic meaning – is normative seems to be flatly false. To be interesting, the idea that word use is normative involves norms that are not reducible to mere causal regularities; that’s the “space or reasons/space of causes” dichotomy that has become de rigeur for many scholars (McDowell 1994; Bakhurst 2011; Huemer 2006). But that requires categorical norms, and not merely conditional norms, for the latter are easily handled in a teleological reduction. However, the idea that linguistic meaning imposes categorical norms on our word use is hard to sustain (Hattiangadi 2007). Some writers on Wittgenstein have accepted the point (Boghossian 2003), but most commentators from Baker and Hacker through to more recent studies continue to repeat the claim that meaning is normative without engaging with the substantive literature that has challenged the very coherence of Kripke’s supposed insight (Williams 2010; McGinn 2013; Child 2010; Stern 2004).

  4. 4.

    The idea that learning involves a social model of initiation is incoherent. Assume that learning involves real cognitive development, e.g., the acquisition of new concepts. This is a challenging assumption, one that invites us to consider how learning so conceived is so much as possible. The invitation is to provide an explanation of how there can be a process by which one acquires new concepts. Many Wittgensteinians think the invitation to explain how learning is possible should be avoided (Bakhurst 2011), although some take the invitation seriously (Williams 1994). Williams’ response is instructively clear and well argued.

    Williams endorses a social model of how concept acquisition is possible in which the key ingredient to the model is outsourced to the social. Williams accepts that Wittgenstein’s concept of training is, on its own, an insufficient resource to make learning possible – point (1) above. For Williams, learning = training plus X and her extra ingredient is outsourced to the learner’s teachers, elders, and others. The learner acquires a new concept by being extended the “courtesy” of being treated as having acquired the concept by her “others.” But that simply begs the question and cannot begin to be a coherent model of concept acquisition without an account of what it is about the individual that renders them apt to accept the courtesy extended by the others if they do not already have the concept in question. In short, such social outsourcing of the ingredient that makes learning possible is either incoherent or it collapses into a form of nativism in which the individual already possesses the concept in question.

    The threat of nativism, if one takes the invitation to explain the possibility of learning, is often taken as good reason to avoid the attempt to provide an explanation and to rest content with a description of learning. But the descriptivist strategy fares no better than Williams’ bold attempt at explanation.

    To advocate description in favor of explanation is, in effect, to agree with Baker and Hacker that all explanations of meaning are “internal” or “intra-linguistic.” But that is a form of nativism, for it means that one can only explain/teach the meaning of a word to someone who already has the conceptual space for understanding the meaning. If all teaching (compare all ostension) only works “within language,” then it can only work for those already equipped with the resources for understanding the word. And that’s a disguised nativism. Here are two ways of seeing this point.

    First, assume that learning is acquiring a new concept C by analysis; e.g., it’s introduced as the word that attaches to things that are F, G, and H. For this to work, the conceptual slot for C must already be there in the combination of those concepts that provide the analysis. So it’s not really a new concept, it is simply a new label for a way of thinking that was already available by combining simpler concepts.

    Second, a more subtle version of this would be to introduce “C” by saying “it’s one of those,” or “it’s like this…,” or “it’s similar to these….” These locutions are Wittgenstein’s favored expressions when he is talking about learning, e.g., PI 69: “This and similar things are called ‘games.’” There are two ways of understanding what is on offer here. Either these open-ended expressions pick out concepts that analyze the target concept C, or they provide something less than a conceptual encounter from which the learner must then build the new concept C. The latter would be a model in which Wittgenstein has an answer to the invitation to explain how meaning is possible. It is my preferred reading of Wittgenstein. The former is the descriptivist position, but this is still analysis and it is no better than the analysis of C in terms of a combination of simpler concepts, F, G, and H. The descriptivist has analysis in terms of concepts that do not, until the learning encounter, have clear linguistic labels. But to understand these explanations – “it’s like this…” – one must already have the conceptual space into which these words fit. This is a more sophisticated nativism, but fully compatible with Fodor’s well-known version. The mind has a stock of innate concepts and learning is simply the transaction by which one acquires labels for these concepts. No learning as such, in the sense of acquiring concepts, takes place.


An Individualist Reading of Wittgenstein

If we take the descriptivist approach and eschew the invitation to explain how learning is possible, then we forego any answer to what seems to be a fundamental question:

What differentiates the subject with a capacity to learn (acquire concepts) from those that do not?

This question needs an individual answer. It asks for an account of what makes the individual apt to be a learner, for no matter how much scaffolding from others might support learning and provide important platforms that speed up the process, without an account of the individual’s resource by which they access the social support, the social has nothing to support.

To be clear: the social is important. It is a powerful resource for learning, but it is not the key constitutive element to answering the invitation to explain how learning is possible. That has to be something about the individual. It has to be something about the individual that explains how by giving them less than a conceptual encounter with things – it’s like this, go this way, these and similar things, etc.… – we can provide them enough whereby they come to grasp a new concept. But this means that we need an explanation, an account of how encounters that are less than conceptually shaped encounters can provide the material from which conceptually shaped encounters can arise. This is a staging solution to the invitation to explain how learning is possible. Something of this form has to be available if we are to respond to the invitation to say how learning is possible. The basic form of a staging solution has to be like this: an individual who lacks the conceptual resources for encountering Fs (lacks the resources for encountering Fs in patterns answerable for truth and falsity) has the resources for encountering Fs in patterns that are not answerable to truth and falsity. You might call these non-conceptual patterns, but the label tells us nothing about what sort of patterns they are, other than that they are not answerable to truth and falsity (i.e., they fail the generality constraint constitutive of conceptual content).

Any account of such a form will meet an obvious response from descriptivist readers of Wittgenstein: surely the dualism inherent in a staging solution inherits the problematic divide in the dichotomy between reasons and causes? The conceptual is the normative realm of reason-giving and the non-conceptual the causal realm of brute encounters. Set up the staging solution as outlined and we’ll never get the two to meet; the space of reasons is distinct to the space of causes. But that riposte simply repeats the fiction that Kripke got Wittgenstein right in saying that meaning is normative. As I have suggested above: (a) Wittgenstein nowhere speaks of rules as norms, (b) Good job too! For there is good reason to think that the idea that there are irreducible norms governing how we use words meaningfully is simply false. Wittgenstein’s own discussion of regularities in word use is much more relaxed than the somewhat fevered normativism found in many commentators.

We use words in patterns that are regular. Rules are like garden paths (BT 240, 243). Paths are natural, they are part of our natural history. They are, in many respects, quite unremarkable things. Our fondness for paths is a natural aspect of our being that we share with many creatures; even sheep manifest a sense of belonging to paths in their heftedness to their pastures. Paths are a natural feature of how we are in the world. They are things we follow with a sense of allegiance that can provide a feeling that we are being led although we would be hard-pressed to say exactly what leads us. It is certainly not a platonic ideal path that guides us and it is difficult to countenance the bare grass leading us. The source of the feeling that we are following the path is not strictly external at all. Wittgenstein remarks:

When I follow the rule, I do not choose. I follow the rule blindly. (PI §219)

The same applies to the garden path. If we ask, “Why blindly?” the right answer is the one Pears gives. We follow “blindly”

because the constraint comes from within – from our own natures – and not from any external force, and so there is nothing to be seen, and it is even questionable whether what we feel should be called ‘constraint.’ (Pears 2006, p. 29)

Note, this is not the “blindly” that comes from internalizing a social norm (cp. Meredith Williams). This is a use of “blindly” that is part of our natural history of our being the kind of creature that has a sense of allegiance to ways of going that reflects something deep about our nature – the kit with which we confront things that, in themselves, give us no guidance whatsoever. Is there more we can say about this apparent “constraint” from within, something that would help make palatable the outline staging solution sketched above? One option would be to see this constraint as a subjective sense of “ought”; see Ginsborg (2011) on “primitive normativity” and Luntley (2015). We want an account of the individual’s mindedness that has them engaged by patterns that are not conceptual and which give them the resources to acquire new concepts (cf Luntley 2015, 2016 for more).

A clue to how to begin lies in Wittgenstein’s core metaphor for meaningful word use – language games. One of the hardest things to do justice to in reading Wittgenstein is the open-endedness of the regularities in our use of words. It provides a radical occasion-sensitivity to meaningful word use that is rarely fully acknowledged, although Travis (2008) is the key exception here.

The fact that Wittgenstein’s central metaphor for these patterns is “game” can be hard to keep in focus. The concept of game serves many purposes. It says something about how the regularities in word use, however well-formed or flaccid, are in an important sense our regularities. They are not regularities sourced in Platonic abstractions; they are regularities in our uses of words, uses that are natural. The point is clear in a key metaphor for Wittgenstein – the garden path. Regularities are like garden paths (BT 240, 243). They are patterns we respect, to which we have some sense of allegiance, but which lay down no prescription about the way to go. The paths we walk with a clear sense of the way to go are also the paths we create as we go. How can this be? And how does this bear on what we add to training to get learning? Here’s a speculative and tantalizing answer.

The patterns found in games are oftentimes not conceptual patterns. Our earliest games with words (the games we play when we understand very little) are games that pick up on formal patterns of language, patterns of rhythm, rhyme, and repetition. These are the sorts of patterns appreciation of which figures large in our aesthetic experience. These are the patterns we naturally (it is how we are) amend and improvise with. We play with these patterns. Why? Perhaps because we are animals that make and enjoy patterns. We are, first and foremost, animals with imagination and the capacity to make and enjoy, indeed relish, patterns. These are not patterns answerable to truth, but they might become such. And although the sheep responds to patterns in their sense of heftedness to a place (a response we share with them), they do not engage in that distinctively natural human practice of playing with patterns. They lack the imagination to leave the path; their paths do not display the elasticity of ours. So what’s distinctive about us qua learner? The answer is that our paths are natural, but bendy! And they bend because we have the imagination to bend them, they are the patterns we make and shape in the service of the aesthetics of experience. In other words, it is the capacity for play and imagination that characterizes the patterns that provide our first encounter with things where that encounter is not one answerable to truth and falsity. The shape of the non-conceptual is aesthetic. That’s the extra to training that gives the start to learning.

It is instructive to think of this in term of Wittgenstein’s Cartesian inheritance, an individualistic account of our mindedness based on an insight due to Descartes. In the Discourse de la Methode §V, Descartes says that what distinguishes an intelligent, or rational, animal from other beasts is the un-boundedness of his capacity to place himself under the sway of reason. That, of course, sounds like the outcome of learning, a sophisticated end point to education. Understood in the context of Wittgenstein’s naturalism about paths, I suspect it can be the starting point.

Generalize Descartes’s thesis: what distinguishes the animal with a capacity to learn, rather than merely be trained, is the un-boundedness of the animal’s capacity when presented with a pattern to place it under another pattern. Animals that make patterns, that enjoy patterns, and that have a sense of allegiance to patterns respond quite differently to training to the so-called dumb brutes. Animals that relish patterns are subjects with imagination. They are richly resourced individuals. They do not carry the nativist’s burden of a mind full of concepts, but they do carry a basic drive to respond to patterns with play – they make a game of the patterns. They “go on” when told “do it like this…” etc. (cp. PI 208) These phrases do not necessarily express fully conceptual encounters with things; they can be markers for patterns of aesthetic engagement that we can adapt free from the constraint of truth. When playful pattern-makers are subjected to S-R conditioning, the result is quite different to subjecting brutes to S-R conditioning. And when our pattern-making hits the distinctive recalcitrance of that which is non-negotiable (the material rather than the social environment), pattern-makers make first contact with the idea of patterns that are not merely imaginative, but which represent. Then the “do it like this” is answerable to truth. Then you start to move from an aesthetic un-boundedness in patterns to the un-boundedness of reasons that Descartes took as the hallmark of res cogitans. We find Wittgenstein’s naturalism in the res imaginatio. It is part of our natural history (we might as well say res animus). The point is simply this: the subject that responds to training with learning does so because they are creatures of the aesthetic, creatures with an innate capacity and drive to make and live in patterns. Training a res imaginatio means providing affordances that engage their sense of aesthetic patterns. It requires a pedagogy framed by playful encounters with the patterns of the aesthetic.


  1. Bakhurst, D. (2011). The formation of reason. Oxford, UK: Wiley-Blackwell.CrossRefGoogle Scholar
  2. Boghossian, P. (2003). The normativity of content. Philosophical Issues, 13, 31–45.CrossRefGoogle Scholar
  3. Child, W. (2010). Wittgenstein. London: Routledge.Google Scholar
  4. Ginsborg, H. (2011). Rule-following and primitive normativity. Journal of Philosophy, CVIII(5), 227–254.CrossRefGoogle Scholar
  5. Hattiangadi, A. (2007). Oughts and thoughts: Rule following and the normativity of content. Oxford: Oxford University Press.CrossRefGoogle Scholar
  6. Huemer, W. (2006). The transition from causes to norms: Wittgenstein on training. Grazer Philosophiche Studien, 71, 205–225.CrossRefGoogle Scholar
  7. Luntley, M. (2008). Training and learning. Educational Philosophy and Theory., 40(5), 695–711.CrossRefGoogle Scholar
  8. Luntley, M. (2012). Training, training, training: The making of second nature and the roots of Wittgenstein’s pragmatism. European Journal of Pragmatism and American Philosophy. Symposia: Wittgenstein and Pragmatism a Reassessment, IV(2), 88–104.Google Scholar
  9. Luntley, M. (2015). Wittgenstein: Opening investigation. Oxford, UK: Wiley-Blackwell.CrossRefGoogle Scholar
  10. Luntley, M. (2016). Wittgenstein and the path to learning. In M. Peter & J. Stickney (Eds.), A companion to Wittgenstein on education: Pedagogical investigations. Berlin: Springer.Google Scholar
  11. McDowell, J. (1994). Mind and world. Cambridge, MA: Harvard University Press.Google Scholar
  12. McGinn, M. (2013). The Routledge guidebook to the philosophical investigations (2nd ed.). London: Routledge; 1st ed. 1997.Google Scholar
  13. Pears, D. (2006). Paradox and platitude in Wittgenstein’s philosophy. Oxford, UK: Clarendon.CrossRefGoogle Scholar
  14. Smeyers, P., & Burbules, N. (2006). Education as initiation into practices. Educational Theory, 56, 439–449.CrossRefGoogle Scholar
  15. Stern, D. (2004). Wittgenstein’s philosophical investigations: An introduction. Cambridge, UK: Cambridge University Press.CrossRefGoogle Scholar
  16. Stickney, J. (2008). Training and mastery of techniques in Wittgenstein’s later philosophy. Educational Philosophy and Theory, 40(5), 678–694.CrossRefGoogle Scholar
  17. Travis, C. (2008). Occasion-sensitivity. Oxford, UK: Oxford University Press.CrossRefGoogle Scholar
  18. Williams, M. (1994). The significance of learning in Wittgenstein’s later philosophy. Canadian Journal of Philosophy, 24, 173–204.CrossRefGoogle Scholar
  19. Williams, M. (2002). Wittgenstein, mind and meaning: Towards a social conception of mind. London: Routledge.Google Scholar
  20. Williams, M. (2010). Blind obedience: Paradox and learning in the later Wittgenstein. London & New York: Routledge.Google Scholar
  21. Williams, M. (2011). Master and novice in the later Wittgenstein. American Philosophical Quarterly, 48(2), 199–211.Google Scholar

Copyright information

© Springer Science+Business Media Singapore 2017

Authors and Affiliations

  1. 1.Department of PhilosophyUniversity of WarwickCoventryUK