By Ian Hacking
Cambridge University Press, 1983, 287 pages
Review by Michael Beach
In this work, Hacking reviews philosophical thought related to science and technology from the perspective of how scientific and technological ideas do or don’t represent reality. He also shows argument around scientific use of ideas and technology to create reality (intervening). Aside from reviewing the main arguments and philosophers involved on the topics he often interjects his own stands on the issues.
An example of a key philosophical debate is eluded to in a quote by Lakatos. His reading of Popper on knowledge growth stated simply is, “people propose, nature disposes” (114).
Hacking makes a number of comparisons between the philosophical perspectives of Lakatos and others such as Popper, Kuhn, Putnam, and Kant. The key phase is one focus, specifically on how (and if) science progresses. For Lakatos, successive research either progresses a theory, or degenerates it (117). In this way, theories are bolstered or unsupported by empiricist efforts.
Some direct comparison between Kuhn and Putnam allows Hacking to clarify. For instance, while Kuhn speaks of scientific revolution, Putnam is focused more on evolution in terms of knowledge growth through rationality (111). Putnam further muddies the knowledge-growth question through the concepts of reference and extension. One of his arguments, for example, is that a given reference may be understood differently by different people, making the extension, including knowledge growth though experiment, essentially impossible (101). If one accepts this premise, then proposals by people (theories) are not universally understood, nor the disposition of nature as neither the proposition nor the disposition are held in common among scientists.
Putnam’s struggle is with meaning. Hacking denotes that a reference is the meaning, or thing, represented by the word. Sense is more like the connotative understanding of the thing, the reference in question (75). If Putnam questions one’s ability to concur with others on either reference or sense, then his questioning of knowledge growth is understandable. The scientific world seems to get around the difference through the practice of dubbing. Where Lakatos would argue that knowledge growth can only be understood in retrospect (118), Hacking argues in favor of dubbing “new natural kinds” which are “often the result of initial speculations which are gradually articulated into theory and experiment” (82).
Ian Hacking’s work shows a mixed message claiming varying schools of scientific philosophy share common ground, yet differ in fundamental ways, stating how such point-by-point opposition between philosophers only means there is ‘underlying agreement’.
By introduction, Hacking makes a case for ‘common ground’. He shares seven areas where he believes Carnap and Popper, and by extension philosophers of science in general, tend to agree (5). Natural science is the best rational thought. Distinction exists between observation and theory. Knowledge is cumulative. Science has a deductive structure. Science depends on precise language. Unity of science methodology exists in each discipline. Finally, the context of justification differs from the context of discovery.
Despite these unifying assertions, pretty much all the rest of the reading shows an evolution, along with examples of fundamental change of thought. For example, Hacking’s first positivist instinct refers to falsifiability as a ‘variant’ of verification (41), yet early in his work (3) he refers to the divided image of Carnap and Popper as betraying a ‘deeper’ difference. It seems difficult to justify such ‘deeper difference’ with simply being ‘variant’. Difference is variable, on a subjective scale. Qualifying words expose subjective opinion. At times Hacking depicts difference as minor, other times as significant.
Hacking describes schools of thought within his own form of structure; realism vs anti-realism, causal vs anti-causal, theoretical entities vs anti-theoretical entities, and the list continues. A specific example referred to earlier was the divided image of Carnap and Popper. Carnap was in favor of science as verifiable. By this he claimed metaphysics is not science, inductive reasoning should be employed, and there are important meanings in language. Popper, on the other hand, stood for science as falsifiable. By this he argued metaphysics leads to science, deductive reasoning should be employed, and calling meanings and language only ‘scholastic’ (4).
Difference can be understood subjectively by degrees. Hacking seems simultaneously to both emphasize and downplay difference. Readers could easily see downplayed example differences as significant.
Among the topics around speculation and experimentation I found the bridging concept of calculation particularly important. A calculation is a form of modeling. Hacking referenced many ideas of his own and others about meanings of speculation (theory) and experimentation (observation). However, until he addressed the bridging aspect of calculation in the speculation-calculation-experimentation framework, the two seemed somewhat independent. In fact, many of Hacking’s reference philosophers argued specifically a lack of connection between theory and empirical data.
This framework also answered a longstanding question for me. So often in science classes teachers would introduce the idea of constants. These constants were usually attached to the name of a scientist who ‘discovered’ or ‘introduced’ the constant. They never were explained. We were just taught how to incorporate a specific constant into a formula to obtain the answer to a specific scientific process. Hacking explains how a calculation comes about from a need to explain a given observation or experimental data set (artifact, phenomenon). Adding a constant to make a calculation consistently approximate the expected outcome allows science to adopt a theory that adheres to accepted scientific principles. The beauty of such a bridging approach is it also allows for change in both theory and experiment without shifting the calculation. The same calculation can be used to support different theories or outcomes.
The resulting approximation becomes yet another central argument Hacking spends considerable time discussing. If a formula and data from empirical observation consistently approximate theoretical prediction, is that bringing us any closer to truth, or just substantiating a theory that purports to stand for truth? Perhaps the substantiation is merely for a given system generally accepted by the larger scientific community at the time of the speculation-calculation-experimentation linkage.