This post is going to be more like a reading report. It’s hard to have super exciting ideas every week; sooner or later I’ll have to write something mundane. However, as Jeff Berrett (a faculty at LPS and member of my dissertation committee) says: a good paper just needs to elicit one interesting thought in a reader that the reader has not thought of before. I don’t know if this is a good standard to judge papers, but it should be good enough a standard for blogs.
I have been reading a lot on measurement theory lately for my project on psychometrics. It is one of those areas that, after seriously thinking about it for just a little bit, I would immediately realize that “of course it’s a complex subject”, but somehow I have never thought to think about it seriously. As a psychology undergrad, I was taught S. S. Stevens’ four types of scales: ratio, interval, ordinal, nominal. In statistical classes, I learned which techniques are allowed for which types of data. Coincidentally, I learned that Likert-type scales (e.g., on a 1-5 ranging from “mostly disagree” to “neutral” to “mostly agree”; primarily used for attitude measurement such as in my climate survey) was not obviously interval when I had to TA for a sociology class. In my psychology class, the Likert scale was given as the canonical example of an interval scale that wasn’t also ratio. In the sociology class I TAed, the Likert scale was introduced as an ordinal scale without controversy.
In any case, all this is to say that I do have some background knowledge on measurement before my investigation, and yet I was still amazed at how differently things unfolded from the way I would naively expect them to.
mathematics, physics, psychology… or the other way around?
Darrigol, Olivier. “Number and measure: Hermann von Helmholtz at the crossroads of mathematics, physics, and psychology.” Studies in History and Philosophy of Science Part A 34.3 (2003): 515-573.
In this extremely long, detailed, clear, and fascinating paper, Olivier Darrigol outlines a number of theories happening around the time of the publication of Zählen und Messen (counting and measuring), in 1887, by Helmholtz, who is sometimes regarded as one of the founders of measurement theory. Without giving you a summary of this paper, I will only point to things I learned from it that surprised me.
Insofar as we are talking about measurement in the context of “mathematics, physics, and psychology”, one immediate assumption I would make is that we would be talking about using physical apparatus to measure psychological phenomena, resulting in mathematical (i.e. numerical) recordings, such as “the speed of my cognitive processing, as measured by the time it takes for me to press this button, is 400ms”. Statements like this, which was a common target in psychophysical research, express a kind of coordination among three elements: a psychological construct, a physical operation, a number. To say that the coordination measures the psychological component is, I think, to assume that the psychological component is the least “fundamental” or “definitive” of the three, which is how we tend to assume nowadays.
But, of course, things were different. Around the time of Helmholtz’s writing, the European scholarly scene was dominated by Kantian thoughts. Psychology was, in some quite concrete sense, more fundamental than physics, whereas mathematical concepts like numbers and quantity, in virtue of being intangible and abstract, were secondary to both psychology and physics. To be sure, not all thinkers held strong Kantian convictions, but arguments in the form of “numbers couldn’t have such and such properties because that would conflict with our experience of everyday objects” were routinely given and taken seriously.
To give an example of the sort of psychology-mathematics interaction that is unusual today: the mathematician Paul Du Bois-Reymond distinguished cardinal and ordinal numbers by arguing that ordinal numbers measure cardinal numbers.
It is difficult to not see these debates in coherentist terms. The world dictates that experience, physical equipment, and numbers all need to come together in a very specific way. Whichever element is fixed first, the meaning of the other elements can follow. And the disagreement was often about which element should be fixed first. For example, Fechner (of the famous Fechner’s law, which is one of my least favourite concepts in psychology) “regarded it as obvious that the smallest noticeable increment of a sensation S was a constant k independent of the intensity of the sensation” (Darrigol, p.536), and derived a numeric relation between experience and physics accordingly.
operationalize for the sake of agreement
Hardcastle, Gary L. “S.S. Stevens and the Origins of Operationism.” Philosophy of Science 62.3 (1995): 404-424.
Another historic paper I read was about how Stevens’ operationism was not influenced by other similar ideas, such as Bridgman’s operational perspective, Skinner’s behaviourism, and verificationism of the logical positivists. It was an interesting piece for me to read because I rarely read pure history. The goal of this paper was primarily of the form “there exist some misunderstandings about history; let me set us straight”.
In any case, I shall only mention one interesting thing that is thematically related to my previous observation: the primary motivation behind Stevens’ operationism was eradicating disagreement in the lab. The biggest worry for him was “unresolvable disagreement” about which psychological concepts are invoked in the lab. The best way to prevent this was to define concepts in terms of observable behaviours resulting from concrete, executable measurement procedures.
In a sense, measurement is again used as coordination, this time not among concepts, but among researchers.
why should we expect coordination in science?
The question is a curious one, since it seems to have a straightforward answer — we should all agree on truth. But any philosopher of science is aware of the immense complexities associated with the equivocation between science and truth. And even if you are a naive realist about science and believe that 1) science converges to truth uniformly, and 2) we all agree on truth, it’s still unclear that we should expect agreement on approximate truth.
There is an increasing (I think) literature on peer disagreement now, not only in terms of disagreement over propositions, but also disagreement over evidence. My sense is that many people are still uncomfortable with the idea that two equally competent agents can look at the same piece of the world, think very hard, and decide that it has different meanings. Surely at least one of them must be irrational, or that there’s important information left out.
I’m not ready to argue that this cannot be true. I’m merely pointing out that this is a strange assumption.
Here is a completely unsupported sociological hypothesis: “natural philosophy”, or “scholarly thinking”, has traditionally been carried out by an extremely homogenous group — rich white men in the case of Europe. They form tightly knit social groups that place high value in conformity and “apprenticeship”. They all think in exactly the same way, and it had worked for a long time. It’s only natural that the epistemic system developed from this kind of community is not equipped to deal with raw, seemingly unexplained disagreement.
We can see a little bit of this just from the brief history described here. At the time of Helmholtz’s writing, psychology was seen as more solid than mathematics. Part of this was Kantian influence. Part of this might also be just that that seemed obvious to them — everywhere they look, their friends see the world in exactly the same way as they do. I wonder to what extent was Stevens’ worry about disagreement in the lab caused by the “professionalization” of science and the diverse perspectives that followed.
My sociological hypothesis might as well be false. However, I do think that the question of “why do we expect agreement over experience?” is a worthwhile one. In my own work, I prefer to relativize measurement for specific purposes. This perspective gets around the problem of interpretive disagreement to some extent, but it is not what people think about when they think measurement. A measurement is supposed to be a description of some aspect of the world, rather than a theoretical step towards a practical goal. However, if it is in fact the case that people disagree over descriptions of the same world, the question, which is more often avoided than asked, is: what to do, then?
- It might happen after all - May 14, 2023
- Another job market data point - December 17, 2022
- Our place in the fediverse - November 30, 2022
New reader here; great post — I thoroughly enjoyed reading it. In the spirit of raising only one interesting thought, I’d quibble only over your speculative hypothesis about the historical/sociological facts about European academic circles and their relation to (in fact as a proposed explanation for, I take it) the expectation of coordination in natural philosophy. Now, there’s no denying that historically what we call “natural philosophers” tend to be a group of extremely homogenous individuals (with a few exceptions; Spinoza and Du Chatelet come to mind). But I think the claim that they “all think in exactly the same way”–or even a weaker claim, perhaps, that they all think pretty much the same way–is simply not true.
Newton and Descartes, for example, had roughly the same observational data at hand (in terms of epistemic access) regarding planetary motions, yet they came to drastically different conclusions about how the data should be accounted for. There’s no compromise; no operationalized concepts for the sake of agreement or anything like that. In fact, their disagreement ran so deep that they had virtually opposing views not only on motion but also on space and time. The famous Leibniz-Clarke Correspondence might, in that regard, be considered a continuation of that sort of deep, ontological disagreement. Psychologists in later centuries and today have faced a rather different question, of course, not least because psychological constructs are often among the most elusive; and while many tend to avoid asking the important “what to do” question you raised at the end, there are also–and have been–serious thinkers who do tackle that very question (like Newton and Descartes did; Skinner and Chomsky come to mind). But scientists’ general expectation for coordination, speaking at a fairly high level of abstraction, is perhaps nothing beyond an unreflective expectation for laws of nature (or Humean regularities, if you prefer), and I’d expect that to hold true psychologically, regardless of academic circles being homogenous or not. And one reason why we don’t see more theoretical disagreements being discussed head-on (here I’m explicitly excluding the “uninteresting”, or “easy”, disagreements, simply due to different experimental paradigms, for example) is perhaps these disagreements run so deep–indeed, so philosophical–that (most) scientists tend to want to dismiss (again, regardless of whether they think alike to begin with or not).
Thanks so much for this! This is a very interesting point. As much as I’m not at all historically inclined, I have always been fascinated by the Leibniz-Clark correspondence, for exactly the reason you outlined. I also agree with your later assessment that the apparent consensus in science today probably hides substantive disagreement that is too deeply entrenched to be noticed. (which is why science needs philosophy!)
There is something admirable about the Leibniz-Clark correspondence, and other similar head-on, no-compromise debates like ones in the “objections and replies” section of Descartes’ Meditations. However, there is sometimes a tendency (possibly a legacy of the logical positivists) to understand these debates as “metaphysical”, or even “merely metaphysical”. The assumption usually is that disagreement is caused by the non-observables — the untestable theoretical assumptions lurking in the background — and “empirical science”, insofar as it is about observables, should not admit irreconcilable disagreement of this sort. Newton’s claim of “I make no hypothesis” can be seen as a separation of this kind.
All in all, I did not mean to say that we shouldn’t aim for coordination, only that we don’t have good reasons to “expect” it straightforwardly. With regard to this, I mostly have the modern “rational agent” model in mind. The idea there seems to be that every agent, insofar as they are rational, should understand the same piece of the world in the same way, with the implication that, if they don’t understand the world in the same way, either they are not fully rational or they don’t have the same pieces of information after all. I find this assumption ungrounded and problematic.
As to whether my sociological hypothesis is true, I truly cannot say.