Agassi’s “Sensationalism” and Popper on the Empirical Basis

1. Introduction

When I was a student at the London School of Economics from 1966–1971, I took courses in the philosophy department. I recall Joseph Agassi’s “Sensationalism” (Agassi 1966) as having created quite a stir. For while its concerns were much wider than this, it included a critique of Popper’s views about basic statements. This drew attention as Agassi was identified with a “Popperian” approach to philosophy. In this paper, I will explain what I take these views of Popper’s about basic statements to be (and what seems to me their distinctive and important character). I will then report on Agassi’s criticism of Popper, and I will, in turn, criticize it. I will conclude by suggesting a possible modification of Popper’s views which I believe would offer a solution to a problem which Agassi’s paper implicitly poses for Popper.

The problem with which I will be particularly concerned relates to what Agassi discusses as “Boyle’s rule” (Agassi 1966: 17):

…whenever a hypothesis and a report of a repeatable experiment contradict each other it is the hypothesis which has to be thrown overboard

More specifically, I will try to respond to the question: why, if we depart from sensationalism as Agassi describes it (Agassi 1966: 1) (i.e. “all our knowledge of the world comes to us through the senses”), should we give epistemological priority to reports about the results of repeatable experiments?

The move that I will make is possibly in a direction that Agassi would favour, because of the emphasis that he has placed on the importance of metaphysics for science. ¹ However, it is not one that Popper would have favoured when he wrote his Logik der Forschung (Popper 1934), as he was, at that time, trying as far as possible to avoid his approach to science depending upon metaphysics. For, as Popper has explained, he did not, at that time, have a theory about how metaphysical ideas could be rationally appraised (Popper, 1972, note 9 to chapter 2): ²

In my Logik der Forschung… [while] I describe[d] myself as a metaphysical realist, I wrongly identified the limits of science with those of arguability. I later changed my mind and argued that non-testable (i.e. irrefutable) metaphysical theories may be rationally arguable

2. Popper and the “Empirical Basis”

I recall Lakatos as having referred to “Boyle’s Rule” in his lectures at the L.S.E. But at the time, I knew too little to draw anything significant from such discussion as took place. On reflection, it seems to me that critical discussion of Popper’s ideas about “the empirical basis” did not receive the attention that they deserved among members of the L.S.E. philosophy department.

More recently, I have become increasingly impressed with what I take Popper’s discussion of “the empirical basis” to amount to. For Popper takes an inter-subjective and realist approach to the problem of how our theories should be assessed. Inter-subjectivity is developed by Popper by reference to themes in Kant’s work. ³ In particular, Popper refers to the Kantian idea that “scientific knowledge should be justifiable ⁴ independently of anybody’s whim: a justification is ‘objective’ if in principle it can be tested and understood by anybody”. (Popper 1959: 44). Popper continues by quoting Kant: “If something is valid… for anybody in possession of his reason, then its grounds are objective and sufficient.” ⁵

On Popper’s account, claims are made by people as to what the results are when, say, an experiment is conducted. They needed, however, also to give an account of how they proceeded, so that the claim that they made about what was to be observed when a particular experiment takes place, could be tested by other people. For Popper, a basic statement is, paradigmatically, ⁶ a tentative claim about what there is, out there, in the world: about that was observed to occur when an observation was made. It is, emphatically, not a claim about our “sense data” or our observational experiences.

The realism of Popper’s approach – emphasising tentative claims about what is to be observed as going on in the world when we undertake tests, rather than our subjective experience about what is going on - seems to me to mark a decisive break with the empiricist tradition. I have found it strange that some in other respects sympathetic critics of Popper – such as (Watkins 1984) and (Zahar 1989; 1995) – have tried to push Popper back towards forms of subjectivist empiricism. ⁷ While Gunnar Andersson’s treatment in his “The Problem of the Empirical Basis in Critical Rationalism” (Andersson 2016) seems to me to favour a similar approach. ⁸

In my view, any such tendency is misguided. Not only does it seem to me that Popper’s view is an attractive account of the empirical basis, but that, if one takes seriously such parallels as Popper sometimes draws between his epistemology and ethics – e.g. in the 1961 Addendum “Facts, Standards and Truth: A Further Criticism of Relativism” to The Open Society (Popper 1966) - one should explore whether his critique of the significance of subjective certainty in his epistemology in (Popper 1959, section 8) can also be applied to moral philosophy. If we were to take such an approach, this should, I believe, lead us to reject Popper’s own tendency to appeal, in ethics, to the judgements of the individual’s conscience. ⁹ However, to make any such move in ethics opens up many more issues than I can address here. ¹⁰ But it would seem to me in line with the broadly objectivist approach to ethics that Popper takes in some of his work. ¹¹

I have mentioned this issue here, just in order to indicate why Popper’s anti-subjectivist approach to basic statements seems to me so interesting and suggestive. Indeed, it could be applied in any field in which we are interested in claims to objective knowledge. It is for this reason that I would like to say more about it here – and to defend it from Agassi’s criticism.

3. What is Involved in Popper’s Approach?

Just so that we have what is involved clearly before us, let me offer a brief summary of what I would take Popper’s view to be.

(a) We are concerned, here, with a (fallible) account of what we take features of reality to be, against which our theoretical claims can be tested.

(b) Popper wrote: (Popper 1959, section 29) “Experiences can motivate a decision [about these things] and hence an acceptance or a rejection of a statement, but a basic statement cannot be justified by them – no more than by thumping the table.”

(c) The content of a basic statement is a report on what was observed when a test took place: it is a statement about what is out there in the world.

(c) But inter-subjective acceptability plays a key role, and reports on basic statements need to be given together with information as to how they are to be tested by others.

(Popper 1959, section 30) refers to basic statements as being “accepted as the result of a decision or agreement” – and then immediately goes on to make what I would take to be a bad move, in comparing this to “conventions”. I would suggest that what is needed here, is simply to see if those who are competent to appraise what is claimed to be the case, raise questions about what are offered as basic statements, and what, then, happens if the issues that they raise are put to the test. That is to say, what is involved is inter-subjective dialogue about what, in fact, there is, out there in the world.

When, earlier, in this section, I referred to Popper’s idea that “experiences can motivate a decision” about a basic statement, I think that we need to take the “can” seriously (i.e. as suggesting that they don’t have to be involved). In Britain, radio broadcasting on Radio 4 concludes, late at night, with a detailed weather report for those at sea. It includes information about what is going on, drawing on reports from a mixture of observers and automatic weather stations. This information is consolidated, along with theoretical information about expected changes in barometric pressure. And, presumably, if there are significant discrepancies between different specific reports, critical investigation takes place before the broadcast is made.

I would suggest that this also offers us a useful picture of what we are involved in more generally, in respect of “basic statements”. They may be motivated by people’s experiences. But they may, surely, be generated by other sources. Further, we have little idea as to what other people’s subjective experiences are like. If questions are to be raised about what they are reporting, we will not be interested in issues concerning the relationship between the statement that they made about what is out there in the world, and a subjective experience which might have prompted them to make it (compare Popper 1963, “Introduction”). Rather, we will wish to look at its coherence with other reports – whether made by people, or by automatic weather stations! In addition, its coherence with our current tentative theoretical knowledge will also play a key role as to whether or not we question the reports: we will be interested in their cogency, not in what people’s subjective experiences were.

What we have, here, looks to me to amount to a coherence approach to epistemology (but, emphatically, not a coherence theory of truth). It is directed towards our aim of discovering interesting truth about the world, by way of inter-subjective appraisal of claims that are made about, inter alia, what is the state of things, when someone observes it. If this is right, it is an utterly different approach to that of sensationalism. For what matters with regard to “basic statements” is whether someone was undertaking an experiment, or reporting on an observation. And whether what they said about what they saw passed the critical scrutiny of others or their own internalization of what is involved in that kind of appraisal. (That is to say, we, as individuals, typically make checks – e.g. before we report on what we have seen – that we have internalized, based on the kinds of critical scrutiny that we have come to receive from others.)

4. Back to Agassi’s “Sensationalism”

It is with this in mind, that I turn back to (Agassi 1966). Much in it – e.g. the survey of the “Sensationalist” tradition – seemed to me interesting, but not something about which I wish, here, to raise any questions.

By contrast, what I found strange was:

(a) Agassi’s view that Popper was in some sense a “sensationalist”.

(b) His criticism of Popper in respect of “Boyle’s rule”.

(c) That it did not seem clear what he was offering as an alternative to Popper’s approach.

I will discuss these points in turn.

(a) I am not sure that, in respect of my first point, I need to do any more than refer to the account that I have already given, above. Popper’s approach – centred upon statements about what is out there in the world, and on the significance of their inter-subjective acceptability – seems to me to break with the sensationalist tradition.

(b) I am puzzled by some of what Agassi says about “Boyle’s Rule”. From a Popperian standpoint, basic statements are, obviously, fallible – not least because they are open to ongoing inter-subjective appraisal. At any point, it is possible for someone to open up the question of whether a particular report about what had been observed was not, in fact, incorrect, and to propose that it should be re-tested. (We would, however, typically expect that, if someone was proposing this, they would explain why they thought it made sense to question something which had, to that point, been accepted.) And we can, clearly, also be led to question hitherto accepted “basic statements” as a result of developments in our theoretical ideas. ¹²

There is, however, something further that in my view needs to be said about these issues. For one might say that Popper’s approach takes over, uncritically, from the empiricist tradition, a view of the role to be played by basic statements. That is to say, someone might indeed ask: why, if there is a clash between a theory and a basic statement, should – other things being equal – the theory be called into question?

That we should do so, seems to me perfectly reasonable. But if Agassi’s argument was that why people did so, was an uncritical residue of the “sensationalist” tradition (which I would agree with him should be repudiated), we need, if this practise is questioned, to offer an alternative rationale for it. To put my point another way, if we see, with J. S. Mill or with Mach, human knowledge as a kind of construction out of sensations; or if, more broadly with the empiricist tradition, we see “impressions”, “sensations”, or sense-data as, in some sense, the foundation of knowledge, then it would seem clear enough why one should accord them priority over our theories, in the event of a clash. But if one parts company with this tradition, then, it seems to me, an alternative explanation needs to be offered as to why (a fallibilist version of) Boyle’s law makes sense.

This, it seems to me, calls for a response in terms of a conjectural non-justificationist metaphysics. This is something at which I can only gesture here, for reasons of space. It is something with which Popper would not have been happy at the time that he wrote (Popper 1934), for the reasons that I explained earlier in the paper. It seems to me, provided that we do not claim that it can serve as a justification for what we are doing, it would be a good move to spell out a tentative picture of what we see as being our situation in the world, of how our perceptual system and our social institutions work, of a kind which would make sense of the sort of critical preference that we give to basic statements in our system of knowledge. The point of doing this, would, on the one hand, be to render coherent the kind of preference that we give to basic statements. On the other, if we explicate such an account, we may be able then to criticise it: not to say that we cannot justify a claim that it is true, but, rather, to suggest ways in which it might be corrected. And such an exercise, and dialogue about it, might, in addition, suggest ways in which our procedures with regard to basic statements might, themselves, be improved.

I offer this tentative suggestion with some trepidation. In part because of its brevity: it is, as it stands, no more than a programmatic suggestion. In part, because I am well aware that it might be very badly faulty. But also because my mind is drawn back to a conversation with the late Bill Bartley about Robert Nozick’s Philosophical Explanations (Nozick, 1981), which I was in the course of reviewing. ¹³ Bartley dismissed the book as “Philosophical Justifications”. But I do not see why a conjectural metaphysics of the kind that I am suggesting here (or, indeed, Nozick’s book) has to be read in this way. Given that we can take a critical approach towards metaphysics, it seems to me that our ideas in this area are enhanced, if we offer other people an account of what, in our view, makes sense of what we are doing, and thus provides them with an object with which they can engage critically. To be sure, our wider metaphysical ideas might turn out to be untenable without this hitting the cogency of our account of basic statements. But if we are in this situation, it should, it seems to me, prompt us to come up with a better metaphysical account – one which we might hope could suggest possible improvements to our account of basic statements, in its turn.