"...the several systematic and topical theories that we retain, at different levels of application, are true to different phenomena and different data domains. Theories are not checked by comparison with a passive world with which we hope they correspond.... We invent devices that produce data and isolate or create phenomena, and a network of different levels of theory is true to these phenomena. Conversely we may in the end count them as phenomena only when the data can be interpreted by theory. Thus there evolves a curious tailor-made fit between our ideas, our apparatus, and our observations." (Hacking 1992, 57-58).
Ian Hacking is known as a strong advocate of entity realism, with the slogan "if you can spray them, then they are real". But the above quoted passages sound much more like anti-realism. Has he converted to anti-realism, or is the quotation compatible with entity realism? The purpose of this paper is to establish a plausible interpretation of Hacking's position.
Hacking's paper's purpose is to answer the question how laboratory sciences become stable despite the fact that so many elements of laboratory experiments are modifiable. His answer is that the symbiosis (mutual self-vindication) of theories and laboratory equipment enables the stability. The above quotation appears as an answer to the question of how we are dealing with the disunity of science as a natural consequence of the self-vindicating nature of laboratory sciences.
The passages sound anti-realistic for the following reasons. First of all, the second sentence ("theories are not checked by ...") is a clear denial of theory realism, and this is consistent with his earlier position. But he also seems to coem out against entity realism. His vocabulary is restricted to the observable world -- "theory", "idea", "phenomena", "data", "apparatus" and "observations", and there is no mention of real entities out there. The stability of laboratory sciences is attained just by adjusting these items with each other. Even phenomena do not have a priority on theory, and what we do is just to evolve a "tailor-made fit". The word "tailor-made" suggests that he thinks that the resulting coherence is artificial. Thus the aim of laboratory sciences is not to account for real entities but to save the tailor-made fit. This is an anti-realist position.
Then, is Hacking an anti-realist now? This conclusion is too hasty. To start with, let us look at his treatment of a "target" (46). A target is "substance or population to be studied" (ibid.) and as such a part of "materiel of experiment" (ibid.). A target is altered or interfered with by a source of modification in experiment. Apparently he is talking about what he called "entity" in his 1983 book. Therefore he still maintains his entity realism. On the other hand, here a target is just one of fifteen elements of experiment he enumerates, and he also adds a rather remarkable qualification: "In the case of Atwood's machine we have neither target nor source of modification; it is a detector pure and simple. There is nothing ultimate about my classification" (ibid.). In other words, a target no longer enjoy the privileged status found in the 1983 book, and it is not even an indispensable part of experimentation. This observation suggests that even though Hacking is still an entity realist, this does not play an important role in this paper. Thus he says: "The thesis has almost nothing to do with recent manifestations of scientific realism or anti-realism, being compatible with almost all the significant assertions made by either party" (31).
Now his position is clear. The next question is how this position fits with the above quoted anti-realistic view. I think that Hacking's following comment is illuminative: "One difference between my 1981 self and Latour is that I did not think of electrons being created, but did think of the photoelectric effect being created, in a pure state" (37). We should be careful because it is stated in the past tense. Does Hacking mean that this is no longer his position? This is unlikely, as is evident in the following sentence: "I am glad that Latour's criticism has enabled me to reiterate one of my favorite themes, the creation of phenomena, previously left out of this essay" (ibid.). Therefore we can take the contrast -- creation of phenomena and non-creation of entity -- as his current position.
This explains why Hacking thinks that the stability of laboratory sciences "has almost nothing to do with" the realism vs. anti-realism debate. Laboratory sciences deal with created phenomena, and the phenomena are sometimes caused by interfering with a target, but not necessarily so. Even when a target is involved, the phenomena caused by the target are not independent of us in the way the target is.
Let us go back to our first quotation. I said that "entities" do not appear here, but maybe they do. For "apparatus" may include a target, and as we saw, "target" is another name for entities. This reading is supported by Hacking's restatement of his thesis: "We modify, I have said, any or all of my fifteen elements in order to bring them into some kind of consilience" (58). A target is, of course, one of the fifteen elements. Thus entities play a role in the adjusting process (if relevant experiments use targets). This is compatible with the artificiality of the stability expressed by the word "tailor-made fit", because of the gap between created phenomena and entities. Therefore the role played by entities (targets) in the process is a small replaceable piece, the replacement of which may create new phenomena. With this interpretation, the quotation fits into Hacking's consistent position.