A few weeks ago I led a session of Coffee with Scientists in Cambridge, on the subject of Harry Collins’ 1985 publication Changing Order, and in particular, his ideas about tacit knowledge. I have blogged about Collins before here, but on that occasion what I said was more of a gut reaction than a measured response. Leading a discussion is a wonderful motivator for actually putting together a reasoned argument, so I thought it appropriate that, having done that, I should now revisit the subject.
Tacit knowledge and the Experimenters’ Regress were two of the big ideas that emerged from Collins’ book. Leaving aside the Experimenters’ Regress until such time as I can bear to discuss it, I decided to take on tacit knowledge because of the number of times I had heard it referred to in seminars and lectures in a seemingly uncritical way. Since my own reaction on reading about it had been somewhat negative, it struck me that either those who have swallowed this doctrine whole have missed something, or else I have.
Collins acknowedges that the concept of tacit knowledge was first articulated by Michael Polanyi, but it is his own work that is often cited in connection with it. He relies rather heavily – unwisely, in my opinion – on the example of riding a bicycle, about which he says, reasonably enough, “no amount of reading and study in the physics and dynamics of the bicycle will enable a novice to get on and ride immediately. On the other hand, the skilled rider is usually quite unable to describe the dynamics of balance involved.”
The case study Collins uses in this chapter (chapter 3) is of attempts to build an atmospheric-pressure carbon dioxide laser (dubbed the TEA, or Transversely Excited Atmospheric-pressure, laser for short) in the 1970s. He communicated with several labs that were trying to replicate the achievements of a Canadian lab that had successfully built one, and actually worked closely with one of the scientists in two separate attempts to build one. His basic thesis is that such exercises cannot be simply reduced to writing down a set of instructions, and the missing component is the direct experience of someone who has already built one, communicated by some other means: tacit knowledge.
My reaction to this chapter was that, while tacit knowledge undoubtedly exists, Collins seems to be greatly exaggerating its importance, in an attempt to surround scientific practice with an aura of mystique which is entirely unjustified. Here are the points I made in the seminar to back up this conclusion.
1. No Cycling Here? Collins does not actually identify any aspect of scientific practice that is anything like riding a bike. He could have mentioned soldering and glassblowing, although such tasks are usually done by technicians rather than scientists, and whilst, like riding a bike, they are basically motor skills, they are generic, so that once someone has learnt how to do them they can apply them to any situation. One tacit skill he would not have been able to cite in the 1970s, but is very much with us now, is operating a computer; but note that many simple computer operations can be written down if necessary; it’s just that we don’t usually do this, and also, if asked how we do something, we often have to first watch ourselves mechanically performing the requisite actions before we can write them down! Maybe we should call such skills “semi-tacit knowledge”? (A participant in the Cambridge seminar suggested that there is actually some real tacit knowledge lurking in the background here, but it relates to knowing how to look for a particular operation, rather than knowing how to do it).
2. Documentation. In order to stress the tacit nature of the knowledge required, Collins claims that “no scientist succeeded in building a laser by using only information found in published or other written sources” (p55). Sorry Harry, but I don’t believe this – or rather, I can believe the particular point you are making about the TEA laser, but I don’t accept that it is not possible in general to find the full specification written down somewhere. Of course one would not expect to find full information in a published paper; for one thing, that would make most scientific papers much too long, and a journal full of them would weigh a ton. But there are usually auxiliary documents which do specify all the nuts and bolts – such documents usually go under a name like “Technical Design Report”. There are of course special circumstances which might have made finding such documentation hard for the TEA laser: at the time Collins became involved with it it had only recently been declassified, and also there was a fair amount of competition between the various labs to build the first one. So nobody was giving anything away, and he does acknowledge that. But that is not the normal situation in physics, and one cannot help wondering why, given that, Collins chose that particular experiment – he doesn’t tell us in the book.
3. Danger – sociologists at work. Collins doesn’t tell us how the scientists he observed reacted to being observed. I am no sociologist, but surely this is an important factor in such situations? One particular scientist, Bob Harrison, invited Collins into his lab and let him help with the experiment. They seem to have got on very well; but mightn’t even Harrison sometimes have got fed up with this? And might that not have resulted in Harrison perhaps providing less than comprehensive answers to all the questions? At one point, Collins tells us “After lunch H left me to work with two graduate students” (p67). He doesn’t say why … maybe he was fed up with being observed? Collins does tell us that “scientists are resistant to the sort of account of experimentation that I have just given”, which could be interpreted as evidence of a clash of some sort.
4. Scientific background. Was Collins’ previous scientific knowledge sufficient for him to be able to understand what was happening in the lab? He doesn’t tell us explicitly, but we get a hint of where he is at scientifically from the following comment: “He then tried all the components of the electronics with an ‘Avometer’, a device that would check that resistances and connections were correct” (p66). This is a bit like someone observing a cleaner at work and commenting that “he cleaned the floor with an Electrolux” – he does not seem to have appreciated that the appropriate term to use here is the generic term for the use to which the multimeter was being put (ohm-meter or continuity tester) rather than the make. Surely if one is going to observe anyone at work one needs a certain basic minimum of understanding of that work?
5. Written records. Was Harrison keeping a record of his work in a lab book? Lab books are a very basic element of laboratory procedure. Yet on page 68, Collins reveals a difference of opinion between him and Harrison regarding whether the power supply voltage had fallen when the spark gap was fired; they seem to be relying on memory here, rather than written notes. A lab in which results are not being written down is hardly a model of good practice.
6. Was Harrison working alone? If he was, it is hardly surprising that he did not have specialist knowledge of such things as high voltage circuit design. He was, we are told, an “expert in non-linear optics”. If he did not know his high-voltage theory well, could he not have consulted someone who did, or even collaborated with them?
7. Was Harrison actually a very good scientist? He seems easily frustrated and impatient, lacking the methodical, dogged qualities that experimental scientists need. For example, Collins quotes him as saying that he was “fairly desperate” (p60) and “there’s a limit to how much checking of the system you are prepared to do before trying it out” (p75). On p69 we learn that he was only interested in the end product and what it could do for him, not in the laser per se: “I wanted lasers as quickly as possible … to be able to get on with a lot of the research which we’ve been trying to do …” It’s also noteworthy that on the occasion when Harrison was absent from the lab, the first thing Collins and the two students did was “check out the characteristics of Jumbo” (p69). A good bit of methodical plodding – that’s the spirit! In fact at this point Collins, who, despite his apparent lack of scientific knowledge, may have had the better temperament, seems to gain the ascendancy over the tired, frustrated Harrison. The work described at the foot of page 67 is a bit of textbook methodical fault-finding.
Harrison is an ideas man, a blue-sky thinker, and building the laser – tried and tested technology – is an engineering job. The two do not always match very well. Happily, over the decades since the Manhattan Project, modern “big science” has gradually embraced engineers, and even though the scientists do not always see eye to eye with them, the engineers do their best to make sure everything is documented and done methodically. So I doubt whether anything like Collins’s experience with the laser could happen nowadays. (But then, as an ex-engineer turned scientist, I would say that, wouldn’t I?)
The above are all, I believe, relevant factors which should be taken into account when assessing whether Collins’ experiences in the laser labs really back up his assertions. Not surprisingly, I would argue that in fact these assertions are not justified. Science studies people delight in saying (often to scientists) that “it’s more complicated than that” – a phenomenon that crops up repeatedly in this excellent blog piece by Oliver Marsh. Well, this scientist would like to return the favour and say to Professor Collins: “It’s a bit more complicated than that, Harry!”
However, that doesn’t mean I do not have the greatest respect for what Collins was trying to do. And besides, he gets my vote for managing to incorporate in his book references to two of my favourite comedies – The Hitch Hikers’ Guide to the Galaxy, and Monty Python’s Flying Circus. (The “Hungarian phrasebook” sketch he refers to was, I think, actually taken from an earlier precursor to Python, possibly The 1948 Show – but that is a mere detail!)