Tacit knowledge

In my most recent blogs about assessment, I’ve looked at some of the practical problems with assessment criteria.  I think these practical problems are related to two theoretical issues: the nature of human judgment, which I’ve written about here, and tacit knowledge, which is what this post is about. In Michael Polanyi’s phrase, ‘we know more than we can tell’, and we certainly know more than we can tell in APP grids.

Take writing as an example. Teachers know what quality writing is, and when given examples of writing, teachers tend to agree on the relative quality of the examples. But it is fiendishly difficult to articulate exactly what makes one piece of writing better quality than another, and still harder to generate a set of rules which will allow a novice to identify or create quality. Sets of rules for creating quality writing or quality anything can descend into absurdity. Dylan Wiliam is fond of quoting the following from Michael Polanyi:

Maxims are rules, the correct application of which is part of the art which they govern. The true maxims of golfing or of poetry increase our insight into golfing or poetry and may even give valuable guidance to golfers and poets; but these maxims would instantly condemn themselves to absurdity if they tried to replace the golfer’s skill or the poet’s art. Maxims cannot be understood, still less applied by anyone not already possessing a good practical knowledge of the art. They derive their interest from our appreciation of the art and cannot themselves either replace or establish that appreciation.

‘These maxims would instantly condemn themselves to absurdity.’ This phrase goes through my mind whenever I read essays that have been self-consciously written to the rules of a mark scheme, rubric, or other kind of maxim. For example, I often read essays which have been quite obviously written to the rules of a PEE paragraph structure.

In this poem, the poet is angry. I know he is angry because it says the word ‘anger’. This shows me that he is angry.


In this extract, Dickens shows us that Magwitch is frightening. I know this because it says ‘bleak’ and this word shows me that Magwitch is very intimidating.

Or, at GCSE, (and this is derived from an examiners’ report, here)

This article tells us that horse-racing is dangerous. We know it is dangerous because it is dangerous.

Or, at A-level, I have read essays where pupils repeat chunks of the assessment objectives, as if to flag up to the examiner that they are ticking this particular objective.

In The Darkling Thrush, Hardy uses an unusual form to shape meaning. He also uses a different structure and his language is very interesting, and overall, the form, structure and language shape meaning in this literary text.

Or, more commonly at primary, writing where every sentence begins with an adverbial word or phrase which barely makes any sense.

Forgettably, he crept through the darkness.

I think the absurdity here results from pupils having been given a rule or maxim which is of some help but which will not on its own create quality. Generally, it is a good idea to use evidence and explain your reasoning, to comment on form, structure and language, and to use adverbial sentence openers. But without concrete examples of how such rules operate in practice, they are of very limited value. And this is the problem with criteria and rubrics: they are full of prose descriptions of what quality is, but they will not actually help anyone who doesn’t already know what quality is to acquire it. Or, in Rob Coe’s words, criteria ‘are not meaningful unless you know what they already mean.’

I’ve argued before that over-reliance on criteria leads to confusion and inaccuracies with grading. But what we see here is even worse: reliance on criteria also leads to confusion in the classroom. Prose criteria are only helpful if you already understand the subject, so using them as a method to inculcate understanding is futile. And yet, we’re often recommended to share ‘success criteria’ with pupils, and rubrics are often rewritten in ‘pupil-friendly language’ which may be pupil friendly in that pupils can pronounce them, but are certainly not pupil friendly in that they can understand what they mean. This approach also leads to a ‘tick-box’ mentality, where pupils and teachers look to make sure that pupils have ticked off everything on the mark scheme. But again, this is unhelpful, because for something like writing, the question is not whether a pupil has used a an adverbial opener or referred to historical context, but is more about how well they have used the adverbial opener, and how appropriate and insightful their reference to context is. The people for whom rubrics and criteria will be least helpful are novices: pupils and new teachers. And yet the people who end up relying on them the most, and who are often encouraged to rely on them as a means to acquire expertise, are pupils and new teachers. Rubrics on their own will not help them to acquire expertise, and in many cases, I worry that they may even inhibit the development of expertise.

Polanyi’s student Thomas Kuhn wrote about the problem of tacit knowledge in The Structure of Scientific Revolutions.

A phenomenon familiar to both students of science and historians of science provides a clue. The former regularly report that they have read through a chapter of their text, understood it perfectly, but nonetheless had difficulty solving a number of the problems at the chapter’s end…learning is not acquired by exclusively verbal means. Rather it comes as one is given words together with concrete examples of how they function in use; nature and words are learned together.

Kuhn is talking about science here; to adapt this for writing, we might say that examples and words are learned together. As I’ve argued here, it is not enough to provide descriptors describing quality writing: descriptors need to be accompanied with examples of what essays of this particular standard look like.

This idea of tacit knowledge can sometimes be interpreted to mean that pupils can never learn something explicitly and must just pick up expertise implicitly. That is not my interpretation at all, nor do I think it is borne out by Kuhn or Polanyi’s work. Kuhn does say that rules and prose descriptions on their own cannot bring understanding, but he does not suggest replacing them with aimless discovery. His suggestion is that it is the problem sets at the end of the prose textbook chapter which really bring meaning. The types of problem sets he is referring to are often quite artificial and isolated examples of the natural world, examples that have been deliberately isolated and selected to prove the textbook’s point. Polanyi also talks quite extensively about how the expert has spent hours focussing their attention on tiny details, and learning to recognise differences that completely elude the casual observer. This is not achieved through discovery, but through direction. It is not achieved quickly, but through thousands of hours of deliberate practice.

Similarly, to go back to the example of writing, I don’t think that we can just expect pupils to pick up notions of quality writing through discovery. What we need are examples of quality writing where the salient features are isolated and discussed, and where pupils have to respond in some way to them, just as the problem sets in a science textbook require certain responses.

If we want to explain what quality is, we need more than just prose descriptors. We need annotated examples, problem sets, and plenty of time.


14 thoughts on “Tacit knowledge

  1. Martin Robinson

    The judgement of what quality is should be part of the conversational encounter in the classroom, an argument against relativism and yet an embrace of subjectivity tempered by tradition… An idea of taste, eloquence, beauty and the sublime. We need to teach aesthetics…

  2. Crispin Weston

    Which is another way of saying what I have been arguing, that we should think of learning objectives as resolving to capabilities – i.e. the dispositions to produce certain sorts of performance. The meaning of such capabilities may well be best communicated by reference to exemplars, but the need to deal with Plato’s problem of multiplicity means that we must be able to share a conception of the abstraction, beyond the ability to point to one or more concrete instantiations of that abstraction.

    The extent to which this abstraction needs to be described explicitly will vary, I suggest, depending on the level at which you are operating.

    1. The guys who are selecting the exemplars need to be able to describe the criteria in detail, so that they can contest the selection of the exemplars and so avoid falling into the “group think” of which Juliet warns in a previous post.

    2. The guys who are marking the student work need to be able to recognise the standards of performance in the sort of visceral sense that Martin describes – a sense which is re-enforced by familiarity with the exemplars – but they do not necessarily need to define the elephant which they can recognise reliably when they see it.

    3. The student does not need to describe the abstract capability at all – they just need to possess it, and they demonstrate this fact by producing the right sorts of performance in the appropriate circumstances.

  3. mmiweb

    I agree that there are implicit dangers in the sharing of “success criteria” and when exemplification becomes the rule. However I would contest that this is the root of the problem, a quick search for P-E-E or other “rules” which show you that they are inevitable linked to statements such as, “how to get a grade higher” or “improving your exam results”. When we have an instrumentalist system of assessment then we more likely get an instrumentalist form of teaching esp. when this is then linked to pay and performance measures – the aesthetics that Martin wants will do.

    The über-abserdity of this comes when there start to be definitions of what is an exclamation and what is a question as set out in the new frameworks for testing (https://michaelt1979.wordpress.com/2015/06/29/when-is-a-question-not-a-question/) so expression, quality and diversity will disappear quickly.

    Finally I would like to argue not for the dichotomy of discovery or direction or explicit or implicit but for a, to coin a phrase, third way. Yes of course there will be a need for direction both in terms of the sharing of the expertise of the teacher and in the teacher offering examples of what they (and perhaps the wider community) think of as excellence but there should also be the opportunity for the pupil to experiment, explore and discover for themselves.

  4. Robert Craigen

    This reminds me of a discussion I had with ministry officials here where I was trying to make the case that the curriculum and teacher resources appear to conflate wisdom for teachers and pedagogical ideas with learning outcomes for students. Asked for an example, off the top of my head, I said “well, the materials seem to suggest that primary grade students are supposed to know what ‘metacognition’ is and that you should be testing for whether or not they’ve learned this … instead of testing the mathematical skills that supposedly arise from using metacognition. Technical phrases like metacognition, and even the focus and awareness of such, is for teachers, not primary-grade students”. A ministry official in the curriculum branch immediately responded that on the contrary, she hopes that students learn that word, it’s a wonderful word, and it will serve them well. Aherm. Well, I suppose we all think it’s cute when 7-year-olds throw out 5-syllable technical terms at the dinner table. But … is sally able to add fractions? Can she just say ‘metacognition’ … or does she actually pay attention when solving a problem to whether or not she’s used all the information in the problem …? Seems to me it’s the same problem Daisy is highlighting here.

    1. The Wing to Heaven Post author

      Robert, I think you are absolutely right. Performance descriptors really quickly
      turn into jargon. The ‘I can’ statements that were really popular in English schools led to pupils parroting statements about things they actually were not able to do and didn’t really understand. It reminds me of Gramsci’s line about crystallising everything in complexity. Performance descriptors actually make learning much more mysterious and difficult than it needs to be.

  5. Pingback: On tacit knowledge Daisy Christodoulou could learn, a thing or two, from Jacques Derrida. | Education: the sacred and the profane

  6. Dick Schutz

    “Tacit knowledge” begs the question of how to best report National achievement test results. That is, by definition, “tacit” is a matter that is “not expressed” and test results are a matter that inherently “are expressed.” We encounter the same complexity in communicating test results that is encountered in instruction. In both situations the audience varies greatly– from naive novices e.g. the students who are being instructed, to sophisticated specialists e.g. psychometricians versed in statistics. Ironically, the novices have more “tacit knowledge” about the communication than the specialists. As Craigen notes, the novices’ concern is not about the experts’ “knowledge,” “higher order skills,” and other matters the experts consider important. Ironically, as yet, the experts are short on the “problem solving skills” and “critical thinking” to bridge the communication gap.

  7. rick nelson (@rnelson696)

    On tacit learning, it may be helpful to distinguish between well-structured and ill-structured problems and domains. In well-structured (rule-based) domains, our brains can apply very complex rules quickly and accurately without being able to say what they are. We all have been able to do this in speech in our native language since we were 3 years of age. How many of us know the rules of grammar? Not many, but yet we all speak following the rules of our dialect quite consistently: in pronunciation, morphology, syntax, and semantics.
    I can solve algebra problems and recognize incorrect algebra, but I am very poor at explaining what “rule” I am using or what rule a student is violating. But for me as a science teacher, speech and algebra are tools. Only linguists and math majors need to be able to state the rules. The great majority of us need only to apply them correctly to do useful work in our fields.
    That is why the requirement in “some interpretations” of the US common core standards that students “explain why” is so frightening. It takes immense time to BOTH be fluent in applying the rules based on implicit, tacit, intuitive but quite accurate rule understanding AND to gain the expert’s ability to “explain why,” What most students need to do productive work is to use speech and math correctly. Too much time on “why” means less fluency, and the way a brain that evolved to support learning fluent speech solves all well-structured problems is by the fluent recall and application of facts (words) and rules. Too much time on “why” and you lack the fluency to solve problems.

  8. Pingback: Comparative judgment: 21st century assessment | The Wing to Heaven

  9. Pingback: Refining assessment without levels | Teaching: Leading Learning

  10. Pingback: Should teachers be told how to teach by those who’ve never been teachers themselves? | chronotope

  11. Mark Bennet

    The two maxims “Many hands make light work.” and “Too many cooks spoil the broth.” have always seemed to me to illustrate a key point – they are a part of our collective wisdom, but it takes experience to know which one to apply. Wise people have often accumulated lots of rules, but there is more to their wisdom than the rule book.


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s