Coherence and Correspondence

Pontius Pilate’s question ‘What is truth?’ addressed to Jesus, of course, invokes an irony to readers of the Gospel of St John. A few chapters earlier, Jesus had told his disciples ‘I am the way, and the truth, and the light’ (John 14:6). At his trial, Pilate inquires of Jesus, the prisoner, ‘What is truth?’ The seeker asks the Truth the nature of truth.

Without getting into the complexities of how the truth could be embodied in a person, we can describe the truth in two general theoretical ways. The first is the correspondence theory and the second is coherence theory. The correspondence theory suggests that truth is not mind-dependent in any significant way. If I see a field, a field is before me. True propositions (‘I see a field’) correspond with reality (there is a field which I am looking at). There are, of course, many nuances of this theory which philosophers and others have delighted in engaging with, but for the moment we can say that the correspondence theory is a form of realism: what you see is what is there.

The coherence theory suggests that what is true coheres with other propositions. A true proposition is in agreement with many other justified true propositions. Those propositions, of course, are coherent themselves with other justified true propositions, and so on. The problem here is that all the coherent propositions might be wrong; the mental structure could be entirely coherent within itself but bear no relationship to the world.

Which of the theories is right? Well, it depends on what you mean but I suspect that, practically, both have elements of correctness in them. We do perceive and all of us believe that most of our perceptions are correct. I can perceive my cat and stroke her. I see her, feel her fur, hear her purring and so on, and these perceptions go together to make the proposition ‘this is my cat’ which I accept as true. I do not, unless I am writing like this, actually think like that. I recognise my cat (and she recognises me) and we interact. There is some correspondence of perception with the world; there is some coherence of perception with other justified truths. We do not really seem to have advanced very far.

Ideas of truth rapidly run into other, equally contested, ideas. Justification is one of them, as is belief. One idea of knowledge is that it is a justified true belief. That is, if I believe something that is both justified and true then I know something. Unfortunately, this idea was debunked by Gettier (Gettier, E. L. (1963) 'Is Justified True Belief Knowledge?', Analysis, vol. 23, no. 6, pp 121-123.) who showed that some, at least, justified true beliefs are not knowledge. Knowledge, then, is another contested concept.

These questions become rather more acute when we look at something that assumes knowledge, truth and justification. For example, an experiment in physics is built on a whole pile of other experiments, theoretical constructs and the relationship between the two. It is often said that theory is experiment laden and experiment is theory-laden, and that is true. The implications of that are far more significant than usually imagined, however.

Suppose I, as I used to, want to carry out an experiment in the spectroscopy of a plasma. This is just a jargon-ridden way of saying I want to look at the light emitted when I put a current through a low-pressure gas, helium in my case. The current causes the atoms to ionise, and the ions collapse into a thin pencil in the centre of the tube. As they have extra energy, the atoms and ions emit light at energies (frequencies) which are known from non-plasma experiments and from theoretical calculations. The shape and location of these spectral lines can be measured and this yields information about how the ions and atoms are affected by being in a plasma.

The experiment, then, rests on a large quantity of assumptions and beliefs. Atomic theory is one of them, which tells me where the spectral lines should be and also goes towards explaining the broadening and shift of them. But there are more theoretical and experimental items in the pile supporting this experiment – ideas about electrical currents and ionisation; about the detection systems; theories of how spectrometers work (the equipment used to spread the light out in wavelength so we can see the spectral lines); measurements of the calibration of the equipment; statistics about the repeatability of the results and so on.

The point here is not the details of the experiments I conducted (you have to be an enthusiast for that), but the mound of previous experiment, theory and testimony about those theories and experiments that there is. The testimony comes in the form of scientific articles that are peer-reviewed and published; in textbooks; in lectures, seminars, over coffee and by working with others in the same laboratory, learning how stuff is done. When the pile of stuff needed to understand a single experiment is considered, it starts to seem a bit remarkable that we can achieve anything like knowledge. No wonder that the claims of science often seem to be very cautious and modest. They have to be.

Truth, in a scientific context, is elusive. I cannot tell you what is true about an atom in a plasma. I can tell you something about how the light from that atom (and a large number of its colleagues) might be like, and what that may tell us about how the atoms are behaving in those conditions. I have a model, partly mathematical, partly in my head, of how these things work. But it is a model, it is not a description of how these things work. Often we forget that our scientific models are models, not descriptions of reality. Is a scientific (or any other sort of) model ‘right’?

The best a scientific model can do, really, is to predict some stuff (say the light emitted by a plasma) and we can see whether the prediction and what we measure match. Further, we can try to explain why they might not. I once received a note on an undergraduate laboratory experiment report praising me for my ability to ‘explain (away) the results’. As the foregoing suggests, I am still trying to work through the implications of that.