2750 words

Introduction

What I term “the Berka-Nash measurement objection” is—I think—one of the most powerful arguments against not only the concept of IQ “measurement” but against psychological “measurement” as a whole—this also compliments my irreducibility of the mental arguments. (Although there are of course contemporary authors who argue that IQ—and other psychological traits—are immeasurable, the Berka-Nash measurement objection I think touches the heart of the matter extremely well). The argument that Karel Berka (1983) mounted in Measurement: Its Concepts, Theories, and Problems is a masterclass in defining what “measurement” means and the rules needed for what designates X is a true measure and Y as a true measurement device. Then Roy Nash (1990) in Intelligence and Realism: A Materialist Critique of IQ brought Berka’s critique of extraphysical (mental) measurement to a broader audience, simplifying some of the concepts that Berka discussed and likened it to the IQ debate, arguing that there is no true property that IQ tests measure, therefore IQ tests aren’t a measurement device and IQ isn’t a measure.

I have found only one response to this critique of mental measurement by hereditarians—that of Brand et al (2003). Brand et al think they have shown that Berka’s and Nash’s critique of mental measurement is consistent with IQ, and that IQ can be seen as a form of “quasi-quantification.” But their response misses the mark. In this article I will argue how it misses the mark and it’s for these reasons: (1) they didn’t articulate the specified measured object, object of measurement and measurement unit for IQ and they overlooked the challenges that Berka discussed about mental measurement; (2) they ignored the lack of objectively reproducible measurement units; (3) they misinterpreted what Berka meant by “quasi-quantification” and then likening it to IQ; and (4) they failed to engage with Berka’s call for precision and reliability.

IQ, therefore, isn’t a measurable construct since there is no property being measured by IQ tests.

Brand et al’s arguments against Berka

The response from Brand et al to Berka’s critiques of mental measurement in the context of IQ raises critical concerns of Berka’s overarching analysis on measurement. So examining their arguments against Berka reveals a few shortcomings which undermine the central tenets of Berka’s thesis of measurement. From failing to articulate the fundamental components of IQ measurement, to overlooking the broader philosophical issues that Berka addressed, Brand et al’s response falls short in providing a comprehensive rebuttal to Berka’s thesis, and in actuality—despite the claims from Brand et al—Berka’s argument against mental measurement doesn’t lend credence to IQ measurement—it effectively destroys it, upon a close, careful reading of Berka (and then Nash).

(1) The lack of articulation of a specified measured object, object of measurement and measurement unit for IQ

This is critical for any claim that X is a measure and that Y is a measurement device—one needs to articulate the specified measured object, object of measurement and measurement unit for what they claim to be measuring. To quote Berka:

“If the necessary preconditions under which the object of measurement can be analyzed on a higher level of qualitative aspects are not satisfied, empirical variables must be related to more concrete equivalence classes of the measured objects. As a rule, we encounter this situation at the very onset of measurement, when it is not yet fully apparent to what sort of objects the property we are searching for refers, when its scope is not precisely delineated, or if we measure it under new conditions which are not entirely clarified operationally and theoretically. This situation is therefore mainly characteristic of the various cases of extra-physical measurement, when it is often not apparent what magnitude is, in fact, measured, or whether that which is measured really corresponds to our projected goals.” (Berka, 1983: 51)

“Both specific postulates of the theory of extraphysical measurement, scaling and testing – the postulates of validity and reliability – are then linked to the thematic area of the meaningfulness of measurement and, to a considerable extent, to the problem area of precision and repeatability. Both these postulates are set forth particularly because the methodologists of extra-physical measurement are very well aware that, unlike in physical measurement, it is here often not at all clear which properties are the actual object of measurement, more precisely, the object of scaling or counting, and what conclusions can be meaningfully derived from the numerical data concerning the assumed subject matter of investigation. Since the formulation, interpretation, and application of these requirements is a subject of very vivid discussion, which so far has not reached any satisfactory and more or less congruent conclusions, in our exposition we shall limit ourselves merely to the most fundamental characteristics of these postulates.” (Berka, 1983: 202-203)

“At any rate, the fact that, in the case of extraphysical measurement, we do not have at our disposal an objectively reproducible and significantly interpretable measurement unit, is the most convincing argument against the conventionalist view of a measurement, as well as against the anti-ontological position of operationalism, instrumentalism, and neopositivism.” (Berka, 1983: 211)

One glaring flaw—and I think it is the biggest—in Brand et al’s response is their failure to articulate the specified measured object, object of measurement and measurement unit for IQ. Berka’s insistence on precision in measurement requires a detailed conception of what IQ tests aim to measure—we know this is “IQ” or “intelligence” or “g“, but they then of course would have run into how to articulate and define it in a physical way. Berka emphasized that the concept of measurement demands precision in defining what is being measured (the specified measured object), the entity being measured (the object of measurement), and the unit applied for measurement (the measurement unit). Thus, for IQ to be a valid measure and for IQ tests to be a valid measurement device, it is crucial to elucidate exactly what the tests measure the nature of the mental attribute which is supposedly under scrutiny, and the standardized unit of measurement.

Berka’s insistence on precision aligns with a fundamental aspect of scientific measurement—the need for a well defined and standardized procedure to quantify a particular property. This is evidence for physical measurement, like the length of an object being measured using meters. But when transitioning to the mental, the challenge lies in actually measuring something that lacks a unit of measurement. (And as Richard Haier (2014) even admits, there is no measurement unit for IQ like inches, liters or grams.) So without a clear and standardized unit for mental properties, claims of measurement are therefore suspect—and impossible. Moreover, by sidestepping this crucial aspect of what Berka was getting at, their argument remains vulnerable to Berka’s foundational challenge regarding the essence of what is being measured along with how it is quantified.

Furthermore, Brand et al failed to grapple with what Berka wrote on mental measurement. Brand et al’s response would have been more robust if it had engaged with Berka’s exploration of the inherent intracacies and nuances involved in establishing a clear object of measurement for IQ, and any mental attributes.

Measurement units have to be a standardized and universally applicable quantity or physical property while allowing for standardized comparisons across different measures. And none exists for IQ, nor any other psychological trait. So we can safely argue that psychometrics isn’t measurement, even without touching contemporary arguments against mental measurement.

(2) Ignoring the lack of objectively reproducible measurement units

A crucial aspect of Berka’s critique involves the absence of objectively reproducible measurement units in the realm of measurement. Berka therefore contended that in the absence of such a standardized unit of measurement, the foundations for a robust enterprise of measurement are compromised. This is yet another thing that Brand et al overlooked in their response.

Brand et al’s response lacks a comprehensive examination of how the absence of objectively reproducible measurement units in mental measurement undermines the claim that IQ is a measure. They do not engage with Berka’s concern that the lack of such units in mental measurement actually hinders the claim that IQ is a measure. So the lack of attention to the absence of objectively reproducible measurement units in mental measurement actually weakens, and I think destroys, Brand et al’s response. They should have explored the ramifications of a so-called measure without a measurement unit. So this then brings me to their claims that IQ is a form of “quasi-quantification.”

(3) Misinterpretation of “quasi-quantification” and its application to IQ

Brand et al hinge their defense of IQ on Berka’s concept of “quasi-quantification”, which they misinterpret. Berka uses “quasi-quantification” to describe situations where the properties being measured lack the clear objectivity and standardization found in actual physical measurements. But Brand et al seem to interpret “quasi-quantification” as a justification for considering IQ as a valid form of measurement.

Brand et al’s misunderstanding of Berka’s conception of “quasi-quantification” is evidence in their attempt to equate it with a validation of IQ as a form of measurement. Berka was not endorsing it as a fully-fledged form of measurement, but he highlighted the limitations and distinctiveness compared to traditional quantification and measurement. Berka distinguishes between quantification, pseudo-quantification, and quasi-quantification. Berka explicitly states that numbering and scaling—in contrast to counting and measurement—cannot be regarded as kinds of quantification. (Note that “counting” in this framework isn’t a variety of measurement, since measurement is much more than enumeration, and counted elements in a set aren’t magnitudes.) Brand et al fail to grasp this nuanced difference, while mischaracterizing quasi-quantification as a blanket acceptance of IQ as a form of measurement.

Berka’s reservations of quasi-quantification are rooted in the challenges and complexities associated with mental properties, acknowledging that they fall short of the clear objectivity found in actual physical measurements. So Brand et al’s interpretation overlooks this critical aspect, which leads them to erroneously argue that accepting IQ as quasi-quantification is sufficient to justify its status as measurement.

Brand et al’s arguments against Nash

Nash’s book, on the other hand, is a much more accessible and pointed attack on the concept of IQ and it’s so-called “measurement.” He spends the book talking about the beginnings of IQ testing to the Flynn Effect, Berka’s argument and then ends with talking about test bias. IQ doesn’t have a true “0” point (like temperature, which IQ-ists have tried to liken to IQ, and the thermometer to IQ tests—there is no lawful property like the relation between mercury and temperature in a thermometer and IQ and intelligence, so again the hereditarian claim fails). But most importantly, Nash made the claim that there is actually no property to be measured by IQ tests—what did he mean by this?

Nash of course doesn’t deny that IQ tests rank individuals on their performance. So the claim that IQ is a metric property is already assumed in IQ theory. But the very fact that people are ranked doesn’t justify the claim that people are then ranked according to a property revealed by their performance (Nash, 1990: 134). Moreover, if intelligence/”IQ” were truly quantifiable, then the difference between 80 and 90 IQ and 110 and 120 IQ would represent the same cognitive difference between both groups of scores. But this isn’t the case.

Nash is a skeptic of the claim that IQ tests measure some property. (As I am.) So he challenges the idea that there is a distinct and quantifiable property that can be objectively measured by IQ tests (the construct “intelligence”). Nash also questions whether intelligence possesses the characteristics necessary for measurement—like a well-defined object of measurement and measurement unit. Nash successfully argued that intelligence cannot be legitimately expressed in a metric concept, since there is no true measurement property. But Brand et al do nothing to attack the arguments of Berka and Nash and they do not at all articulate the specified measured object, object of measurement and measurement unit for IQ, which was the heart of the critique. Furthermore, a precise articulation of the specified measured object when it comes to the metrication of X (any psychological trait) is necessary for the claim that X is a measure (along with articulating the object of measurement and measurement unit). But Brand et al did not address this in their response to Nash, which I think is very telling.

Brand et al do rightly note Nash’s key points, but they fall far, far from the mark in effectively mounting a sound argument against his view. Nash argues that IQ test results can only, at best, be used for ordinal comparisons of “less than, equal to, greater than” (which is also what Michell, 2022 argues, and the concludes the same as Nash). This is of course true, since people take a test and their performance is based on the type of culture they are exposed to (their cultural and psychological tools). Brand et al failed to acknowledge this and grapple with its full implications. But the issue is, Brand et al did not grapple at all with this:

The psychometric literature is full of plaintive appeals that despite all the theoretical difficulties IQ tests must measure something, but we have seen that this is an error. No precise specification of the measured object, no object of measurement, and no measurement unit, means that the necessary conditions for metrication do not exist. (Nash, 1990: 145)

All in all, a fair reading of both Berka and Nash will show that Brand et al slithered away from doing any actual philosophizing on the phenomena that Berka and Nash discussed. And, therefore, that their “response” is anything but.

Conclusion

Berka’s and Nash’s arguments against mental measurement/IQ show the insurmountable challenges that the peddlers of mental measurement have to contend with. Berka emphasized the necessity of clearly defining the measured object, object of measurement and measurement unit for a genuine quantitative measurement—these are the necessary conditions for metrication, and they are nonexistent for IQ. Nash then extended this critique to IQ testing, then concluding that the lack of a measurable property undermines the claim that IQ is a true measurement.

Brand et al’s response, on the other hand, was pitiful. They attempted to reconcile Berka’s concept of “quasi-quantification” with IQ measurement. Despite seemingly having some familiarity with both Berka’s and Nash’s arguments, they did not articulate the specified measured object, object of measurement and measurement unit for IQ. If Berka really did agree that IQ is “quasi-quantification”, then why did Brand et al not articulate what needs to be articulated?

When discussing Nash, Brand et al failed to address Nash’s claim that IQ can only IQ can only allow for ordinal comparisons. Nash emphasized numerous times in his book that an absence of a true measurement property challenges the claim that IQ can be measured. Thus, again, Brand et al’s response did not successfully and effectively engage with Nash’s key points and his overall argument against the possibility of intelligence/IQ (and mental measurement as a whole).

Berka’s and Nash’s critiques highlight the difficulties of treating intelligence (and psychological traits as a whole) as quantifiable properties. Brand et al did not adequately address and consider the issues I brought up above, and they outright tried to weasle their way into having Berka “agree” with them (on quasi-quantification). So they didn’t provide any effective counterargument against them, nor did they do the simplest thing they could have done—which was articulate the specified measured object, object of measurement and measurement unit for IQ. The very fact that there is no true “0” point is devestating for claims that IQ is a measure. I’ve been told on more than one occasion that “IQ is a unit-less measure”—but they doesn’t make sense. That’s just trying to cover for the fact that there is no measurement unit at all, and consequently, no specified measured object and object of measurement.

For these reasons, the Berka-Nash measurement objection remains untouched and the questions raised by them remain unanswered. (It’s simple: IQ-ists just need to admit that they can’t answer the challenge and that psychological traits aren’t measurable like physical traits. But then their whole worldview would crumble.) Maybe we’ll wait another 40 and 30 years for a response to the Berka-Nash measurement objection, and hopefully it will at least try harder than Brand et al did in their failure to address these conceptual issues raised by Berka and Nash.