The Interface between Philosophy of Science and Philosophy of Technology
V. Instrumental Realism
Indiana University Press
Bloomington, 1991, 98-114
At the beginning of this itinerary, I pointed to the different and often divergent interests, concerns, parentages, and traditions associated with two of philosophy’s subdisciplines: the philosophy of science and the philosophy of technology. With the appearance of what I call loosely the “school” of instrumental realists, I have located an important area of interface between these two subdisciplines. Of course, this group of philosophers remains a minority within both of the subdisciplines which continue to be dominated by the habits and concerns of earlier momenta. There is also something of a favored selectivity to this group of thinkers - either they are individuals who have always been recognized as philosophers of science, or if the interest arises out of philosophy of technology, as in my case, it is because their interests are focused in counterpart epistemological and ontological issues.
Yet the very convergence and apparent consensus regarding a critique of dominant momenta and what is shown in the positive analysis of experiment and instrument as the location of one technological embodiment of science is itself of considerable importance. I shall then trace both the broader areas of consensus which have appeared between philosophy of science and philosophy of technology, and also between the Anglo- and Euro-American traditions within which this convergence has occurred. Again, as in all the previous discussions, I will remain somewhat more on the side of the implications for philosophy of science as seen from the contact with a concern for technology than from what remains often more central for philosophers of technology.
First, the consensus regarding a critique of the extant and dominant forms of philosophy of science: I have already pointed out that particularly in North America, the dominant strands of philosophy of science arise from positivist and, later, analytic roots. This has meant that the conception of science and the mode of analysis of science has largely been a matter of logical, linguistic, or propositional analysis. Of course, both the narrowness of that tradition and its tendency to disembody science from many of its concrete dimensions has come under attack - at least since the emergence of the new philosophy of
science. Placing that attack in perspective and relating it to the underappreciated European developments which preceded or paralleled Kuhn was part of the program of the earlier chapters. I have tried to show, moreover, that much of the counterattack directed “irrationalism” and “relativism” continues to miss the mark and that rather, what the dominant philosophies of science have failed to note are the elements of praxis and perception wgich are operative in science. Part of what I take to be the consensus of the instrumental realists is either the direct or the implicit recognition of those elements of scientific activity.
Given the current state of the philosophy of science, with the virtual collapse of the older Positivist strains and the pluralization both through new concerns and the intrusions of sociological and historical revisions, it would be too strong to claim that what I have identified as the school of instrumental realists stands as unique and inflated within the now divergent directions of the New philosophies of science.
This is not even the case with respect to the emergence of what could be called praxis philosophies of science. A healthy development here has been the rise of the sociology of science to its current sophistication. Equally healthy is the emergence of a new generation of bilingual philosophers of science, trained in both Anglo- and Euro-American traditions, who break the older citation gaps earlier referred to. In the next chapter I shall return to this updating of both the moments I have chosen. But here, following the trajectory of praxis-perception philosophies, I wish to further cement the distinctiveness of the instrumental realists.
The focal point at which instrumental realism emerges is the simultaneous recognition of what I have called the technological embodiment of science, which occurs through the instruments and within experimental situations; and of the larger role of praxis and perception through such technologies. It is that focus which locates the positive side of the instrumental realist consensus and which is ultimately more important than the negative critique of extant philosophy of science. Nevertheless, the emergence of a critical consensus from such different starting points of the negative critique is also initially informative.
In its broadest sense, the five philosophers examined here all are conscious and critical of philosophy’s predilection for the “purely theoretical,” often with an accompanying disdain for, or ignorance of, praxis. In general, “armchair” philosophizing is seen as removed from the actual activity of the subject-domain which is being investigated. This predilection is a selective handicap when it comes to either appreciating, or even noticing, the practices of actual science. In this, Dreyfus is the most pointed in critique. He lays the predilection primarily at the door of one strand of tradition - the Platonic-Cartesian - which he thinks has become dominant in science at the
expense of other possible traditions such as the Aristotelian through phenomenological ones. But in all five cases and beyond the most general recognition of the predilection, a much more specific critique emerges: Any philosophy of science which is limited solely to linguistic, logical, or propositional methods will not be able to adequately account for large sectors of scientific activity which entails more than such “rational” procedures alone.
The occasions for this critique - which may be seen as a general challenge to much older philosophy of science - vary among the members of this “school.” Hacking, clearly hostile to older positivism, also wishes to overcome the penchant for “semantic ascent,” which wants only talk about talk concerning such activities as observation and calls for a return to the analysis of observation; I take this to be an Anglo-American equivalent to the Husserlian “things themselves.” Ackermann criticizes the same penchant because it causes dominant philosophy of science to overlook the dynamics of science, thus causing it to fall prey to what he continues to regard as the irrationalist-relativist aspects of the new philosophy of science.
The three phenomenologically oriented philosophers arrive at the same point, although more out of agreement within their own traditions about the role of perception and the embodied position of all epistemological activities noted in phenomenological epistemology. The primacy of praxis and perception is much more explicit within this tradition than in those of the Anglo-Americans.
Beyond the recognition of a general predilection toward theory and the overlooking of praxis, and beyond the general critique of methods linked solely to propositional dimensions of analysis, there is a third aspect within the negative consensus which also may be noted. This aspect arises out of certain results of the positive analysis of the instrumental. I anticipate that result here. By turning to the role of instruments and taking account of what they deliver for science, the very territories previously taken as theoretical domains also change - they shrink in size and significance. In short, instrumental realism gives some degree or type of “reality status” to entities often taken to be merely theoretical, leaving only small areas to remain theoretical. This means, in turn, that the role of “pure” theorizing gets reduced to an even smaller area of science’s activity than had previously been assumed. Theorizing becomes a special, highly speculative exercise of scientific imagination - important, but both reduced in size and open to greater skepticism - in regions outside the current reaches of instrumental possibility. Interestingly, this indirectly implies a more positive evaluation of the results of science than has been the case in either of the parent traditions.
Here, then, we have an emergent consensus regarding a critique of much extant philosophy of science. This critique is arrived at differently
in the two traditions, as would be expected within philosophy: At one extreme we find Dreyfus, who regards a whole strand of tradition to be in error, even considering what Hacking would call the representations of science; but Hacking also arrives at a similar point through his appreciation of a more Baconian science, which both intervenes and represents, and who, through this vision concerning science, wishes to restore a balance between science’s theoretical-representative activity and its intervening activity in experiment. From a slightly different angle, Ackermann wishes to restore the much older faith in progress in scientific knowledge, which he takes to have been under attack since Kuhn. But in the process, he must also attack the narrowness of the tradition which overlooks the dynamics of an instrumentally oriented science. Heelan, perhaps the most optimistic about what instrumentation can deliver, and I, more cautious, find the same results from the way in which instruments embody perception.
The negative critique also has a carry-over to the previous traditions from which both Anglo- and Euro-American thinkers have progressed. If the narrowness and shortsightedness of positivist-analytic philosophy of science, insofar as it restricts itself to propositional analysis, is an object of attack, there are criticisms directed at more direct parentages as well. Such critiques of their own respective traditions cannot be seen except, in part, from the positive results of a newly attained stance. And that stance is taken from the appreciation and analysis of the positive role of instruments within science. I shall shortly turn to that role, but in transition, we may take account of how this shift to the focus upon science’s technologies has effected a critical role within the relevant philosophical traditions.
In each case, the critique carries with it a reevaluation of science itself. From the side of the Anglo-Americans, it is clear that Hacking is restoring something of a balance to what had been a serious imbalance of philosophical concerns. By reestablishing his more Baconian history of science, Hacking calls for and illustrates a more balanced view of science as experimental, and, in that process, necessarily deals with science’s technologies. This reevaluation finds particular result in the reappraisal of observation, which not only becomes a direct concern for Hacking (rather than located through semantic ascent) but is re-placed in its sensory and often instrument-mediated context. In the process, Hacking denies any sharp distinction between “observation” and “theory,” part of the critique of which opens the way to a positive role for instruments. They not only make phenomena present (or manipulable, the favored notion of Hacking) but go on to enhance sensing. Like each of the others, he also recognizes that a “theory bias” frequently distorts the history of science. Hacking then moves away from both a propositional analysis alone and a sharp theory/observation distinction toward the analysis
of the role of instrumentation, which is usually overlooked entirely or underplayed in the dominant tradition.
Ackermann makes a somewhat stronger claim for instrumentation by linking the progress of science itself to the progress of instrumentation. Its history needs rewriting in this light. For Ackermann, the usual distinctions between science and technology need reexamination, and his is clearly an appreciation of even larger aspects of science’s embodiment in technologies. Ackermann comes from a somewhat Popperian heritage which is also strong on the social institutionalization of science, and is itself abstracted out of the dominant traditions.
Ackermann recognizes that in a way parallel to what I shall momentarily point up in the European traditions, even the new philosophy of science lacks appreciation and insight into the role of instrumentation in scientific discovery. Kuhn is criticized by both Ackermann and Hacking for remaining too theory bound and too unappreciative of instruments in “paradigm shifts.”
The same lack of concern for technologies appears within the European traditions. Indeed, the lack of focus upon a technologically embodied science lies at the root of two often noted problems in classical phenomenology. What appears as the wider attack upon theory/observation distinctions also plays a related role in the Husserlian traditions. In their earliest form, these traditions distinguish between the bodily-perceptual aspects of a lifeworld, a position which strongly differs from the “world” of science with its “theoretical” or “ideal” entities. The latter are presumably reduced entities in the Modern philosophical sense in having been reduced to “geometrical-physical” entities lacking sensory dimensions, and they are always derived from the plenary, sensory world of bodily existence.
Heelan points out that Wilfred Sellars has developed a similar distinction (drawn in part from Sellars’s reading of Husserl); but in Sellars’s case, there is an inversion of what is taken to be foundational. Whereas Husserl regards the primary world to be that of bodily-sensoriness, which becomes the “manifest” image in Sellars’s inversion, the “scientific” image, which is the microworld of physics, etc., becomes primary and the base for Sellars’s explanation. I have also pointed out that with an almost, although re-inverted, Sellarsian distinction, Merleau-Ponty continued the Husserlian version of primacy in what he calls the “distance” between the world of science and the lifeworld.
The earlier position of classical phenomenology did make a distinction between these domains in terms of a theory of constitution. Ordinary objects in comparison to scientific objects are differently constituted. But what the classical tradition did not note was that the latter are often, if not typically, instrumentally constituted. Technology –
instrumentation - makes the difference. Lacking the appreciation of the role of instruments which could embody perception, the earlier phenomenological tradition tended to interpret science’s constructs as derivative and abstract, i.e., one could say “merely theoretical,” in comparison to the plenary richness of primary perception. Given this implicit evaluation concerning the comparative reality-status of entities, it was often implied that phenomenology remained negatively predisposed to science.
The insertion of the instrument in a role seen to embody and enhance precisely the bodily (“material”) perceptual intelligence recognized as central in phenomenology was to close this gap between the lifeworld and the world of science. And with it, the residual “negative” characterization of science also could disappear. Here then is a critique of classical phenomenology from the perspective of a more contemporary Euro-American phenomenology, focused specifically on a new appreciation of instrumentation as the embodied mediation needed to make the unification of lifeworld and scientific objects possible. Thus, from within both the Anglo-American and the Euro-American traditions, the instrumental realist consensus arrives at the interface between science and technology - and between the philosophies of science and of technology.
If the critique’s negative dimension is important, this is more in relation to problems within related traditions and to the possibility of reformulations of problems which arise from such perspectives than it is to the more important aspects of the positive dimension of instrumental realism. Thus, I shall now turn to the features of that positive role.
In its broadest sense, the instrumental realist consensus points up the importance of science’s technologies as the means by which discovery occurs and knowledge is expanded. I put the issue in this common sensical form first, even if it is a weaker claim than the one I think more interesting and penetrating, i.e., that contemporary science is more than accidentally - it is essentially - embodied technologically in its instrumentation.
Again, as we might suspect, there is a spectrum of opinion on the role of this technology, albeit focused on the role of knowledge-gathering instrumentation. Hacking’s approach is again the most common sensical and conservative. His argument for the need to focus upon instrumentation seems to be largely historical-empirical. Science has always been Baconian in that Modern science not only represents, but intervenes. Empirically, both representation and intervention are - at least sometimes, and now, increasingly - tied to instruments or technologies. Not to regard this fact is to ignore important aspects of how science actually operates.
Hacking’s call for a re-Baconized concept of science, however, retains its concern for a positive role for science’s representations as
well. What Hacking does, in part, is to say that if philosophers of science want to continue some form of realism, they have been looking in the wrong direction. They would be better off seeing that experimentalists become pragmatic realists in the familiarization of interactions occurring through instrumental manipulations of the entities which become part of the furniture of a scientific world.
What complicates making too strong a claim about an essential role for instrumentation in contemporary science is Hacking’s “pluralism” of methods. He believes in both a plurality of scientific methods and a plurality of experimental approaches. Also, with respect to another issue which divides our five philosophers, he clearly does not hold that all scientific observation is either tied to or occurs exclusively in the presence of instruments.
Yet a closer examination of Hacking’s examples and more particularly, his claims about what kind of activity observation presupposes, reveals that he is at least implicitly much closer to the harder positions taken by others of this “school.” It is abundantly clear - and I shall not cite the many relevant passages which support this - that observation in its scientific sense (a) is a skill, (b) is a matter of trained learning, (c) is active and not passive seeing, and (d) sometimes contains exceptional insight peculiar to technical skills.
Moreover, what the examples show is that (1) almost all observations occur in an instrumental context, (2) there is an implied human-instrument relation, and (3) even what is cited as “sense data” reporting is instrument-mediated and/or constituted. A particularly telling example is his citing of William Herschel’s noticing a phenomenological difference between heat and light transmission:
He had been using coloured filters in one of his telescopes. He noticed that filters of different colors transmit different amounts of heat: “When I used some of them I felt a sensation of heat, though I had but little light, while others gave me much light with scarce any sensation of heat.” We shall not find a better sense-datum report than this, in the whole history of science. 
Herschel went on to use thermometers to test the same phenomenon, finding thereby that “not only orange warms more than indigo, but that there is also a heating effect below the visible red spectrum.”  Herschel then surmised that there were both visible and invisible rays coming from the sun - only much later was this observation appreciated and understood theoretically.
Reverting to my own terminology, I would note that (1) here is an observation thoroughly technologically contexted, (2) the first observation of heat and light differences “sensed” by Herschel were (a) instrumentally mediated and (b) of what I call the embodiment sort,
i.e., perceptually felt. The second set of observations, which used thermometers while testing the “same” phenomenon, were of the hermeneutic sort, both mediated, and since “read,” of a more indirect type than the first set of observations. But both were clearly instrumentally contexted and mediated - and in the stronger sense noted by Heelan, instrumentally “constituted.” The effects were “carpentered” via the prism and lenses
The strong sense of a technologically constituted data-domain, while clearly pointed up by Hacking, is not made with the same force as it is in Ackermann’s, Heelan’s, or my accounts. Yet there is the recognition that many such areas of experimental results are a growing field of observation. These include all those micro- and macrophenomena which lie below or beyond ordinary sensory capacity and which are made present only through or with instruments. At the least, the instrument serves as a condition for such observations; Hacking recognizes and analyzes these conditions in his history of the microscope.
If Hacking is the most cautionary of the philosophers with respect to the essential embodiment of science in technology, Heelan makes the strongest - if also the narrowest - claim to the essentiality of this form of embodiment. For Heelan, as we have noted, only those phenomena which have been instrumentally “carpentered” and “constituted” can have claim to scientific “reality.” This is in part because, for Heelan, all scientific perception is simultaneously perception plus measurement; for measurement to be scientific, it must entail a standardized technology or instrument. Here then is an approach which holds to a necessary connection between scientific observation and its technologies. Of course, one outcome of this position is also that in contrast to Hacking’s pluralism of methods, Heelan comes close to holding that there is a Scientific method.
While such a strong position might appear to narrow what can be taken as the “world” of science because Heelan is also the most optimistic about what realities are exhibited by technologically embodied science, that is not the case. For Heelan, the World is open to both a pre-scientific and a scientific mode of perception. The pre-scientific World is clearly limited to noninstrumentally mediated perceptions, while the scientific world is perceived only through instruments. Adapting a version of Sellars’s “manifest”/”scientific” image distinction, Heelan claims: “A scientific image... represents objects as constituted in their essential forms by systems of postulated (or theoretical) entities, related to one another and to some manifest World by scientific theory, and encountered only through the mediation of instruments or technology. [italics mine]” 
What makes the pre-scientific and the scientific worlds overlap is that both are “perceived”; however, the latter is perceived only through
“readable technologies.” Heelan argues that such a “reading” is direct and equivalent to perception. It is, however, a technology-embedded perception through the “readable technologies” of instrumentation. In this, Heelan continues the physicist’s use of such observations as “direct observation,” a use which Hacking criticizes.” 
Although, from my point of view, Heelan collapses what I distinguish as a difference between embodiment and hermeneutic relations through technologies, his hermeneutic account of “readable technologies” does make of science a specialized mode of world exhibition which, on one side, also “perceives” the world, although only through both formal and technological mediations. Its range of phenomena to be revealed both overlaps the directly perceived world and enhances it, insofar as any entity which is instrumentally detectable becomes a candidate for a reality-status.
Dreyfus occupies a peculiar status within the spectrum of instrumental realists regarding the relation between science and its technological embodiment. That is largely due to the fact that his program is very different from the overlapping ones of the other four philosophers being discussed. On one side, it might seem that Dreyfus is the most negative of the five with respect to the role of technological embodiment of science, and also the closest to the older phenomenological tradition’s suspicion of the role of science. But that is partly because Dreyfus is not directing himself to instrumentation as such, but to a certain interpretation about a special use of machines. His program strongly contrasts machine and human perceptual results. Only indirectly, so far as the instrumental realism at issue here, can the notion of technological embodiment be noted in Dreyfus . But it is there: A subcontention might well be that AI is the attempt to embody a precisely Platonic-Cartesian theory instrumentally (and since this theory is wrong for Dreyfus, it must fail as a model for human embodied intelligence).
Ackermann, too, occupies a fairly strong embodiment position in the science-technology interrelationship. Interestingly, he argues against one traditional “pure theory”/”applied science” argument on the grounds that there may be an ulterior motive to preserving such a distinction by making “pure” theoreticians less open to moral, social, or political criticism (pure knowledge presumably is neutral). Ultimately, Ackermann argues that such a distinction between science and technology has “precluded an understanding of the dynamics of modern science” and that understanding science “is inconceivable apart from that [instrumental] technology.”  Short of Heelan, however, Ackermann sees the dependence of science upon technology as only partial. Technology, nevertheless, provides a major component of modern science in its instrumentarium. He does not, however, argue that only instrumentally constituted entities count as scientific entities.
My position is close to Ackermann’s in terms of placement on the spectrum. I argue that science, in its modern sense, has differed from classical (Greek) science from the beginning in its technological embodiment, both with respect to the actual use of instruments, as in the case of Galileo. His self-conscious use of instruments was proclaimed in “The Heavenly Messenger,” one of the first science newsletters. But such use of technologies was also already part of the Renaissance lifeworld of the time. Modern science is and has been essentially and historically technologically embodied.
Of course, each of our authors notes the obvious: that science is now increasingly so embodied, and necessarily so, with respect to those regions of investigation which fall below or above what is directly sensed in unaided perception.
With respect to what I call a technological embodiment, there remains something of a spectrum of opinion which also relates to the third feature of the consensus binding the instrumental realists together. This, too, is a strongly positive feature of the consensus. For whether science constitutes its entities either partially or solely through technological constructions, one leaves the realm of what is visible to unaided perception in the micro- and macrolevels discovered by science, all clearly agree that enhanced perception occurs only through such instrumentation.
If we momentarily disregard the differences of opinion about whether or not all perceptions of science must be instrument related - within the realm of the micro- and the macrolevels now occupying far larger regions in scientific interest than the “middle-sized” ones of unaided perception—the question of an essential embodiment becomes a matter of agreement. As I noted previously, Heidegger argues that “technology reveals the world” and here we find a new sense for that phrase – technology reveals the micro- and macroworld which lies beyond unaided sense.
Our instrumental realists emphasize this enhancement and magnification made possible through instrumentation. It is here that the narrowing of the region of what has previously been thought of as “theoretical” becomes replaced with the instrumentally “observable,” and in differing degrees, this observability in turn becomes part of a new perceptual region. Here is the heart of the “realism” of instrumental realism. But here too are the areas of most interesting variations upon the broader theme.
On the spectrum of a reality-status for instrumentally delivered entities, Ackermann takes the most conservative position. He argues both for the largest degree of ambiguity relating to what he calls “data domains” which are text-like, and for considerable skepticism relating to the (hermeneutic) interpretive process. “Instrumental means only produce a data text whose relationship to nature is problematic.”  And,
“the features of the world revealed to experiment cannot be philosophically proven to be revealing of the world’s properties.”  (I am not sure whether Ackermann holds any counterpart thesis that other means of analyzing world properties can be philosophically proven!) Yet, examination of the interpretive result relating to such data domains turns out to be the same as for other knowledge claims.
The fact that knowledge progresses and accumulates through instrumental progress, that instruments serve as invariants for multiple observations, and that refinement leads eventually to sufficient agreement all point to the strongest claim Ackermann can make within his dialogical and consensus-driven interpretation of (hermeneutical) science. At some point, further argument - and by extension, skepticism of an extreme kind - become irrelevant.
At the other end of the spectrum, but only in a nuanced way, one again finds Heelan. Within his narrower definition of scientific realities, the instrument delivers most. This is not at all to say that Heelan disregards the ambiguities pointed to by each of the other thinkers in the discussion of instrument-mediated phenomena. He discusses at length precisely the problem of “textual” opacity, ambiguity, etc., which worries the other instrumental realists. But he remains the most optimistic in relation to what is taken to be the “directness” of instrumental mediation and to what I term the collapse of perceptual characteristics into hermeneutic ones. “Reading” becomes a variant upon (direct) perception. It will be seen that both Ackermann and Heelan emphasize a mode of hermeneutic “seeing,” despite the fact that they respectively occupy the two extreme positions on the spectrum of reality-deliverance.
By contrast, Hacking and I occupy the middle positions of this spectrum, with my position somewhat more optimistic than Hacking’s regarding reality-deliverance. Our midpositions are qualified due to concerns over not only what is delivered but the way in which the instrument transforms as well as delivers the phenomenon in question. However, we arrive independently at our concerns for instrumental transformation of phenomena, with Hacking taking what I shall call an externalist approach (inferring what can be seen via current physical theory, as in the application of optical theory to visual result), whereas I take a phenomenological approach utilizing variation and difference from the multiple profiles which occur between instrumental and non-instrumental contexts.
In spite of these differences, there remains consensus regarding the positive delivery of scientifically constituted entities, the role of instrumentation in making this possible, the enhancement of perceptual role, and the subsequent narrowing of the area of “theory” previously thought to lie outside what is (scientifically) experienceable. Yet our four philosophy of science-oriented thinkers all point to a significant
and agreed-upon realm in which there are instrumentally delivered realities. These realities need further attention, but they also constitute a large and interesting realm at the interface of science and of technology.
If the technological embodiment of science becomes the focus for what can be analyzed within this interface area, it is in the domain of praxis-perception taken broadly, but also somewhat thinly, which becomes the last part of the instrumental realist consensus. The praxis part of that model is clearly the stronger element within the consensus. This is because each of our philosophers has taken activity and particularly the instrument and experiment-oriented activity of science more seriously than older dominant forms of analysis. Furthermore, each recognizes the wider, more institutionally oriented ways in which science operates within contemporary Big Science - in a more thorough fashion than earlier positivistic or analytic modes of interpretation. Ackermann’s position is the strongest in this respect, but in each case the others would find a related position. At the very least, both the instrument designer and the skilled users of instruments are raised to higher importance within the overall institution of science.
When it comes to the perception side of the praxis-perception model, I suggest that it might seem our set of philosophers would find the largest areas of disagreement. This may be expected because there is a clearer division here between the Anglo- and the Euro-American groups over what might count as a theory of perception. This is precisely because the latter come from a very strong and more sharply defined tradition regarding perception, in contrast to the vestigially empiricist theories of perception which might be expected of the Anglo-Americans. But that turns out not to be entirely the case. However, there is something of an implicit, contrasted-to-higher-degree explicitness involved here between these two groups.
Through these differences, however, there is also an area which could be called consensual. First, all five of our thinkers notice and agree that sensory-perceptual dimensions of observation are of high importance within the observational side of science. This agreement - other than being taken for granted but often overlooked precisely because so taken - might not appear exceptional, but linked to the more specific recognition of the role of visual imagination, modelling, and even metaphorization, it constitutes a strong sub-domain concerning the way science operates. This visualism is not only recognized in different ways by each of our authors but is held to be an essential part of scientific (perceptual) praxis. Even more strongly, there is mutual recognition that such visualism within science occurs in gestalt and often intuitional ways with respect to insight. What may be called “visual thinking” plays a much larger role than is recognized in most philosophy of science. And this visualism can and does play a role
in and through instrumentation, as well as directly. I would like to argue that this part of the consensus indirectly calls for a careful phenomenological analysis of visual perception. That analysis is in part undertaken by each of the three instrumental realists of the Euro-American side.
A division among our instrumental realists is not provided by the above, but by two different ways of cutting into the role of observation in its perceptual dimensions. And these divisions do not follow Anglo-Euro-American orientations. The first such division relates to whether or not all scientific reality is constituted necessarily through instrumental mediation. At one level, only Heelan occupies an affirmative position on this issue. Each of the other philosophers holds to some variation of observation being possible, either through instrumental or through unmediated perception. While, common-sensically, Heelan’s position may appear extreme, a closer examination shows that it is considerably more complicated, since such a position entails what is taken to be “science,” and subordinately, what can be taken to be a “scientific perception.”
All of our philosophers agree that any scientific perception is not simply an ordinary perception - and not because all scientific perceptions are simply “theory laden.” Rather, our instrumental realists hold diverse views regarding the acute and active skill which must count as part of a scientific perception.
This is to say that there is a peculiar disciplining of perception which is required within science. But in addition to the implied hermeneutics of such a discipline, there are forms of tacit or unspoken skills which many times appear with skilled observers with (or without) instruments.
Hacking, for example, clearly holds that what I call a scientific observation need not be instrumentally mediated, although many are (and their numbers are increasing). But historically, many important observations occur both without instruments (unaided perception) and, more importantly, without an explicit relation to measurements. In contrast, Heelan’s position holds that scientific perceptions are through “readable technologies” and are also perceptions plus measurements. Hacking, like Heelan, recognizes that in more recent science, there are many technologically devised experiments which produce their own carpentered results which, in turn, become highly important to science itself, even if the artifacts of those experiments are not found directly in nature. Indeed, the previously cited example of the laws of thermodynamics arising out of observations concerning the steam engine are interestingly reexamined in precisely that way by Hacking. It might seem that there is a deeper division here than turns out to be the case.
Hacking also points out that the obsession with measurement is historically a late moment in the history of modern science, although
in a stronger sense it does characterize most contemporary science. Granting that measurements have been part of scientific praxis in some sense since even before modern science (among the Babylonians, for example), Hacking points out, “Our conception of numbers and measuring is clear and unquestioned only at the end of the nineteenth century... Kuhn suggests that there was a second scientific revolution, during which a spectrum of physical science is, for the first time, ‘mathematized.’ He puts this somewhere between 1800 and 1850.”  Given what I have called Heelan’s implicit use of physics - one should now say contemporary physics - as the model for science, in contrast to Hacking’s insistence on a pluralism of both theoretical and experimental “sciences,” one sees that the argument is about what shall count as science itself.
However, the issue of necessary technological constitution does not relate simply to measurement. In the context of developing the notion of observation as a skill - often independent of and in advance of theories which can accommodate observations - Hacking notes that:
Caroline Herschel (sister of William) discovered more comets than any other person in history. She got eight in a single year. Several things helped her do this. She was indefatigable. Every moment of cloudless night she was at her station... But most important of all, she could recognize a comet at once... [She] could tell a comet just by looking. 
This skilled, persistent, and gestalt form of scientific perception was not a matter of measurement per se, nor was it something inferred from mathematical projections - although it could have been. Hacking indicates, “Everyone except possibly brother William had to follow the path of the suspected comet before reaching any opinion on its nature.” (Comets have parabolic trajectories.) 
Nor, interestingly, were these brilliant observations initially instrumentally mediated - most of her locating observations were made without the use of even a telescope - but on closer examination, they did occur in an instrumental context in a different way. “She used a device, reconstructed only in 1980 by Michael Hoskin, that enabled her, each night, to scan the entire sky, slice by slice, never skimping on any corner of the heavens.” 
This would appear to be an example which is, at its core, both unmediated and non-measurement oriented, yet clearly a “scientific perception.” Yet while the core of the example does illustrate this, there is also now a recognized instrument-context in both the scanning device and the later telescopic confirmation of initial observations. This example points to a related question, one which focuses more pointedly on the variant understandings of what count for perceptions within the domain of “scientific perceptions.”
Heelan’s notion of scientific perceptions is perhaps more narrow than that of the others, necessarily linked to instrumental contexts; but it is also cast in hermeneutic form. As I have earlier noted, his closest compatriot in this respect is Ackermann. Instruments provide scientific reality through their “readability” and a resultant “data domain,” which is continuously interpreted. The difference between Ackermann and Heelan, at this point, is their understanding of a role of interpretation. Heelan’s greater optimism concerning instrument delivery of scientific “reality” is related to his inclusion of the features of direct perception within his hermeneutic notion of perception.
Within this issue lies the second area of disagreement among our instrumental realists. If the first is to determine whether all, or only some, scientific perceptions are instrument mediated, the second is over the issue of how perception itself is to be construed. It is the issue which lies between a totally hermeneuticized and a body-retained perception.
The issue is not clearly divided between the Anglo- and the Euro-Americans, although it becomes sharper from within the Euro-American tradition than between the two traditions. But first, a brief look at the Anglo-Americans: As already noted, Ackermann’s position is largely hermeneutic - instruments produce data domains which are interpreted. But interpretation is largely a matter of gradual refinement and arrival at a consensus within the institutional context of science. Ackermann, in short, does not often focus upon individualized perceptions. Also lacking is any keen sense of bodily position as an important factor in perception. Instead, he focuses upon what I have earlier called “macroperception,” i.e., perception in its largely cultured sense. Such a perception, of course, is ipso facto a matter of larger hermeneutic concern. And because it is this institutionalized form of perception which plays the larger role in Ackermann, he overlaps Heelan’s hermeneuticizing of perception.
In some contrast with Ackermann, Hacking does draw heavily from historical individuals and upon a version of sense-data tradition which implicitly related to individualized, hence embodied, observers. In that indirect way, Hacking can overlap from a neo-empiricist perspective the emphases of the bodily oriented phenomenologists. Thus, in some areas, Hacking comes closer to both Dreyfus and myself.
Stating the issue as one in which some strength of distinction is to be maintained between a bodily-perceptual (micro-) and a hermeneutic perceptual (macro-) mode, Dreyfus again emerges as having the strongest body-perceptual emphasis. Not only is the human body implied in all its intelligent actions, but its positionality, its gestalt features, and particularly its kinesthetic motility determine its perception of reality.
Not that these elements are ignored by Heelan, the most extreme
of the hermeneutic perceptualists among the Euro-Americans. Indeed, his epistemology quite explicitly recognizes and elevates bodily positionality as a reason for the rejection of both transcendentalism and the implied all-seeing “eye” of Platonic or Cartesian interpreted science. Yet in the end, bodily perception is collapsed into hermeneutic perception. It is the text metaphor which becomes dominant in this version of scientific perception.
This hermeneuticizing of perception may be rephrased by noting that Heelan emphasizes what he takes to be the direct perceptual aspects of reading. “... The value of that quantity [of a scientific entity] can be ‘read’ directly from this response (providing one is experienced and skillful in the use of the instrument).”  Such experienced “readings” are, of course, perceptual experiences. There is a gestalt quality of any form of skilled and already learned reading. And because these perceptual characteristics also characterize the hermeneutic use of instruments, Heelan justifies the collapse of the perceptual aspect into its specialized form of “reading.”
“Reading” temperature is like reading a text... a “text” is “written” causally by the environment under standard circumstances... on the thermometer, and this “text” is “read” as being “about” a presented object, here, the temperature... Such a process is, I claim, essentially both hermeneutical and perceptual. 
Yet, as all the examples show, it is the hermeneutic dimension which absorbs the perceptual one.
If Dreyfus’s position clearly emphasizes the bodily perceptuality of thought and Heelan’s position hermeneuticizes bodily perception, my position falls between these two. As noted before, I distinguish among several modes of human-technology relations which may be phenomenologically analyzed. Of these, embodiment relations and hermeneutic relations are the most relevant to this context. In this distinction, both how one experiences the instrument and the deliverance of the referent make a difference. I may illustrate this difference by drawing from both Hacking’s Herschel siblings and Heelan’s thermometer examples.
Caroline Herschel’s eyeball observations of comets were direct perceptual observations. Clearly, within the context of a highly skilled seeing, and informed by background information on differences between comets and other eyeball astronomical phenomena, she was able to bodily perceive with gestalt instantaneity her selected referents. When an instrument delivers just such a perception, isomorphic though transformed in some way - usually by a reduction of qualities - I call the perception “embodied.” Such is the case with William Herschel’s heat perceptions, mediated through lenses and filters. He felt the radiant
heat upon his hand - and as Hacking went on to describe, he made claims for “measured” felt differences to the thousandths of degrees, which could not have been confirmed by any extant instruments of the time. The phenomenon of heat - transformed and reduced from the fuller presence of focused light - was bodily, perceptually, felt.
By contrast, both Herschel’s totally instrument-mediated and Heelan’s “read” thermometer produce a “text” which refers differently. It does refer, thus it does properly belong to the realm of instrumental realism; but it refers in ways not isomorphic with bodily-sensory relations to objects. Its “textual quality” is presenced in numbers, alphabets, or some other “read” feature which is not isomorphic with that to which it refers. Such hermeneutic relations, however, may also include various manipulable factors sufficient to keep them close to Hacking’s instrumental manipulability of otherwise unsensed entities. In that sense and despite differences, each of the positions remains within the limits noted for an instrumental realism such as has been outlined here.
I shall not further pursue what I take to be both the identifiability of a “school” of instrumental realists and the variations within this “school” which may be areas of fruitful and interesting debate and discussion. Were I to have a wish as an author, I would hope that readers would read, in parallel, the philosophers’ books related to the issues here discussed. It would be my contention that any time such a diverse group of philosophers arrive at such a consensus, often independently and even in some isolation from one another, the issues evolved may well be important ones. The broadest of those issues in the present context centers on the need to more deeply concentrate upon the role of science’s technology, its instrumental embodiment. And this is the interface area which connects philosophy of science and philosophy of technology.