Progress in Science — II

[for a brief explanation of this ongoing series, as well as a full table of contents, go here]

Progress in science: some philosophical accounts

I now turn to some philosophical considerations about progress in science. The literature here is vast, as it encompasses large swaths of epistemology and philosophy of science. Since what you are reading is not a graduate level textbook in philosophy of science, I will focus my remarks primarily on some recent overviews of the subject matter by Niiniluoto (2011, an expansion and update of Niiniluoto 1980) and Bird (2007, 2010), because they capture much of what I think needs to be said for my purposes here. Niiniluoto (2011) in particular will offer the interested reader plenty of additional references to expand one’s understanding on this issue beyond what is required in this chapter. Readers with a more general (i.e., less technical) interest in the history of ideas in philosophy of science should consult the invaluable Chalmers (2013). There are many other excellent sources and interesting viewpoints out there, however, and I will address some of them as needed throughout the remainder of this discussion.

A good starting point for conceptual analyses of progress in science is to trace their roots back a few centuries, particularly the period between the Renaissance and the Enlightenment, when people began to take seriously the idea that “natural philosophy” was a new and potentially very powerful kid on the block when it came to the augmentation of human knowledge and understanding. The simplest view had been held by scientists themselves since at least the 17th century (e.g., Robert Boyle and Robert Hooke), and according to Niiniluoto (1980) can be traced back to the 15th century and Nicholas of Cusa’s concept of “learned ignorance”: science makes progress because it accumulates truths about the world (what I have been calling the “teleonomic” account). But philosophers, beginning as early as the 18th century, pointed out that this assumes a rather optimistic (some would even say naive) view of the epistemic powers of science. Nonetheless, the optimistic attitude endured through the Enlightenment (particularly via Auguste Comte’s positivism) and led to familiar positions, such as this one by Sarton (1936): “progress has no definite and unquestionable meaning in other fields than the field of science.” As we have seen earlier in the book, I know a number of scientists who still happily subscribe without much qualification to Sarton’s take on the matter.

By the 19th century, however, some philosophers were beginning to articulate more nuanced or qualified perspectives, while still maintaining that scientific progress is to be understood as an accumulation of knowledge about the world (and, since according to the standard account of knowledge, the latter is equivalent to justified true beliefs, this means that science accumulates truths about the world) [2]. We have previously encountered, for instance, Charles Peirce’s pragmatic take on truth, which led him to think of it as the limit — in a mathematical sense — of scientific inquiry: “We hope that in the progress of science its error will indefinitely diminish, just as the error of 3.14159, the value given for π, will indefinitely diminish as the calculation is carried to more and more places of decimal” (Peirce, quoted in Niiniluoto, 1980, 432). The analogy, however, is problematic for a variety of reasons, a main one being that it can be convincingly argued that there is no reason to think we have any guarantee of monotonic convergence in scientific knowledge, while there is in mathematical knowledge, at least in the case of relatively simple mathematical problems, such as the calculation with ever increasing degrees of accuracy of the digits of π.

Parenthetically, it is interesting to consider a perennial side discussion related to the idea of scientific progress: if science does make progress, will it eventually come to an end? Science journalist John Horgan (1996) got into trouble when he asked a number of scientists (and philosophers!) that very question, since many scientists apparently just refuse to take it (or the idea of limits to scientific knowledge) seriously. Indeed, according to Niiniluoto (1980, 434), already astronomer John Herschel is on record has having stated, back in 1831: “[the world is an] inexhaustible store which only awaits our continued endeavors,” although George Gore argued the opposite in 1878, on the ground that we will either have solved all problems worth solving, or we will run out of technical and epistemic resources. I tend to agree with Gore rather than Herschel, but the point is that the very question implies some account of how science actually progresses. If there is no progress, then it is harder to see in which sense one can meaningfully ask if the process will reach an end.

The debate on how exactly science makes progress really took off in philosophy of science in the ‘60s and ’70, particularly with the contributions of Popper, Kuhn, and Feyerabend. Setting aside the latter’s “radical” epistemology (according to which there is no such thing as scientific methodology, anything goes as long as it delivers whatever results we are after), a major distinction can be drawn between Popper’s and Kuhn’s views (e.g, Rowbottom 2011). In Popper’s view scientific progress results from a continuous approximation to the truth, achieved via falsification of previously held theories. This was part of Popper’s well known attempt at overcoming Hume’s problem of induction, which led him to rethink the scientific approach in terms of falsification rather than confirmation (because confirmation of theories is too easy to achieve, and does not even separate science from pseudoscience: Pigliucci and Boudry 2013). Popper’s views, however, run afoul of the well known Duhem-Quine problem, which we have already discussed.

Kuhn famously framed the issue of progress in science in a more neutral fashion. Indeed, so neutral that he was quickly accused of advancing a framework that makes it impossible to actually talk about scientific progress, lending him accusations of relativism, which he vigorously rejected late in his career (Kuhn 1982). Kuhn’s view was that we can easily make sense of progress within a given paradigm: during the so-called “puzzle solving” phase of scientific discovery scientists deploy the conceptual and instrumental tools made available by the reigning paradigm to solve a number of local problems (“puzzles”). The more problems are thus solved, the more one can say science is making (local) progress. According to Kuhn, however, at some point there will be a sufficiently high number of unsolved puzzles, which will lead to a crisis of confidence in the paradigm and eventually to its replacement by a new paradigm. What is difficult to say — and even Kuhn himself had a hard time articulating it — is in what sense moving from one paradigm to the other counts as progress. In fact, Kuhn’s famous analogy between paradigms and gestalt switches in the psychology of perception (those cases where the same image can be interpreted in completely different ways by the brain) did not help, since the two alternative perceptions of a gestalt image have equal claim to be the “correct” interpretation (worse: neither one, technically, is correct since the images are designed on purpose in order to be ambiguous). Kuhn really did sound at some point as if he were saying that paradigms are convenient frameworks for doing science, with no way to determine whether and in what sense a given paradigm may be better than another one.

Moreover, things get apparently worse for a Kuhnian view of scientific progress because of the existence of what are usually referred to as “Kuhn losses.” These are instances where puzzles that were solved under the old paradigm reemerge as problematic under the new one. For instance, the old phlogiston theory in physics accounted for why metals have similar properties (they all contain phlogiston). Except that it turned out that there is no such thing as phlogiston, so the problem was reopened (and solved again, in terms of modern atomic theory). Kuhn losses open up the possibility of scientific regress, at least locally. A number of issues immediately present themselves once we start looking at scientific progress this way. To begin with, what exactly counts as a “problem” or “puzzle”? Depending on how we answer this question — which at the very least is bound to be specific to subfields of the natural sciences — our estimate of Kuhn losses may vary dramatically. Also, as much as the Kuhnian broad view can be taken to be neutral with respect to the issue of progress in science, certainly it will be difficult to talk about solving puzzles without any reference to concepts such as truth or truth-likeness, so that again we arrive at least at a minimalist view of progress. Ultimately, it is an issue for historians of science to determine the relative frequency of Kuhn losses, and it may turn out to be the case that their numbers are much smaller compared to the number of new puzzles that are solved after a paradigm shift, so that a meaningful — even quantifiable — sense of progress in science could be recovered even within a Kuhnian framework.

Kuhn himself became quickly aware of these issues, and attempted to articulate a positive view of progress in science in his famous Postscript to the last edition of The Structure of Scientific Revolutions. There he does three things to clarify his position and respond to his critics: he argues that philosophy of science is both prescriptive and descriptive, so that any accusation that he mixed up the two roles is beside the point. He also spent a significant amount of time elaborating on his central concept of “paradigm,” re-defining it as a disciplinary matrix that includes not just whatever dominant scientific theory holds the field in a given area of inquiry (e.g., Newtonian mechanics, or general relativity), but also the ensemble of accepted experimental and analytical methods, ancillary concepts and hypotheses, what counts as relevant or important questions that remain to be addressed, and even the type of training for graduate and undergraduate students, which is the way the new generation of scientists is introduced to the dominant paradigm. Crucially for my discussion here, however, what Kuhn also attempted in the Postscript was a defense of the idea of progress in science. That defense was only partial: he used the metaphor of an evolutionary tree of scientific ideas and suggested that science progresses along the unfolding of new theories, which branch out of old ones. But he also admitted that he was a “relativist” in the narrow sense that he didn’t believe that scientists can meaningfully talk about reality “out there” in a theory-independent fashion. Given its importance and influence on all subsequent discourse, it is worth quoting that passage in full here:

“Imagine an evolutionary tree representing the development of the modern scientific specialties from their common origins in, say, primitive natural philosophy and the crafts. A line drawn up that tree, never doubling back, from the trunk to the tip of some branch would trace a succession of theories related by descent. Considering any two such theories, chosen from points not too near their origin, it should be easy to design a list of criteria that would enable an uncommitted observer to distinguish the earlier from the more recent theory time after time. Among the most useful would be: accuracy of prediction, particularly of quantitative prediction; the balance between esoteric and everyday subject matter; and the number of different problems solved. Less useful for this purpose, though also important determinants of scientific life, would be such values as simplicity, scope, and compatibility with other specialties. Those lists are not yet the ones required, but I have no doubt that they can be completed. If they can, then scientific development is, like biological, a unidirectional and irreversible process. Later scientific theories are better than earlier ones for solving puzzles in the often quite different environments to which they are applied. That is not a relativist’s position, and it displays the sense in which I am a convinced believer in scientific progress.”

“Compared with the notion of progress most prevalent among both philosophers of science and laymen, however, this position lacks an essential element. A scientific theory is usually felt to be better than its predecessors not only in the sense that it is a better instrument for discovering and solving puzzles but also because it is somehow a better representation of what nature is really like. One often hears that successive theories grow ever closer to, or approximate more and more closely to, the truth. Apparently generalisations like that refer not to the puzzle-solutions and the concrete predictions derived from a theory but rather to its ontology, to the match, that is, between the entities with which the theory populates nature and what is ‘really there.’”

“Perhaps there is some other way of salvaging the notion of ‘truth’ for application to whole theories, but this one will not do. There is, I think, no theory-independent way to reconstruct phrases like ‘really there’; the notion of a match between the ontology of a theory and its ‘real’ counterpart in nature now seems to me illusive in principle. Besides, as a historian, I am impressed with the implausibility of the view. I do not doubt, for example, that Newton’s mechanics improves on Aristotle’s and that Einstein’s improves on Newton’s as instruments for puzzle-solving. But I can see in their succession no coherent direction of ontological development. On the contrary, in some important respects, though by no means in all, Einstein’s general theory of relativity is closer to Aristotle’s than either of them is to Newton’s. Though the temptation to describe that position as relativistic is understandable, the description seems to me wrong. Conversely, if the position be relativism, I cannot see that the relativist loses anything needed to account for the nature and development of the sciences.” (Kuhn, 2012, 204-205; my italics)

It is in the context of both Popper and Kuhn that our next quick entry makes sense: Imre Lakatos’ (1963/64, 1970) idea of scientific research programmes. Lakatos (a student of Popper) sought to overcome the opposition between what he saw as the logicist approach put forth by his mentor and the more psychological take elaborated by Kuhn, while retaining advantages of both. He therefore suggested that science does make progress, but via what he called research programmes. These are the Lakatosian equivalent of Kuhn’s paradigms, and therefore much broader than the specific theories that Popper discussed in terms of falsification. Research programs are made of a “hard core” and a “protective belt”: the first is comprised of whatever theoretical commitment would, if abandoned, essentially spell the end of the programme itself; the second one includes “expendable” ancillary hypotheses or theoretical constructs which, if abandoned, would not trigger a Kuhnian crisis. For instance, part of the hard core of the Copernican theory was the idea that the Earth revolves around the Sun, not the other way around. However, Copernicus’ initial assumption that the orbits of the planets are circular was part of the protective belt. When that particular idea was abandoned, by Kepler — who realized that the orbits must be elliptical instead — the theory survived, and the programme kept being, in Lakatos’ terminology, “progressive,” meaning that it led to further research and discovery. Nevertheless, sometimes research programmes do run into significant problems, and have to rely on their protective belt in an increasingly ad hoc fashion, precisely the sort of thing that Popper would have said would doom any serious scientific theory. Before Copernicus, for instance, the Ptolemaic system had to be supplemented with an increasing number of (entirely imaginary, as it turned out) epicycles in order to keep it in reasonable, though still inaccurate, working order.

Thanks to the Popper-Kuhn-Lakatos debates of the 1960s and ‘70s, then, philosophers of science arrived at an understanding of a number of ways in which science can be said to be progressive, from the more cautious one put forth by Kuhn above, to the more Popperian (in spirit) version elaborated by Lakatos. Let me now jump to some of the more recent literature (Niiniluoto 2011), where a distinction is made between progress of science in the axiological (i.e., normative) sense and more neutral descriptions, such as “development.” To simplify a bit, the basic idea is that scientists and philosophers are concerned with the axiology of the scientific enterprise, while sociologists and historians of science tend to take a descriptive approach, although it seems that no project aiming at understanding such a complex set of issues can do without some congruence between descriptive and prescriptive undertakings — just like Kuhn had suggested in his Postscript.

We can usefully conceptualize much contemporary philosophical discourse on scientific progress following a simple classification proposed by Bird (2007): progress can be thought of in an epistemic, functionalist, or semantic key. From an epistemic perspective (the one that Bird himself favors), scientific progress is cashed out as accumulation of (scientific) knowledge; functionally, progress is defined in terms of a particular function, as in the case of Kuhn’s (and, later, Larry Laudan’s 1981a) problem solving ability, as we have already seen; semantically, progress is an issue of increasing verisimilitude, as was proposed by Popper and has been articulated in more detail recently by Niiniluoto. We have already explored the functionalist approach to some extent, so I will now focus on Bird’s epistemic take and Niiniluoto’s semantic one, though discussing both of them will require going back to Kuhn-Laudan style functionalism as a useful contrast.

Bird’s epistemic account assumes the standard (if controversial among epistemologists) definition of knowledge as justified true belief, of which scientific knowledge is a particular subset. That is because Bird — correctly — wants to avoid counting as knowledge instances in which one arrives at the right conclusion by sheer luck. In this sense, Bird simply continues a long tradition in philosophy of science according to which the discipline is concerned with the context of justification of scientific theorizing, as opposed to the context of scientific discovery, which is best left to the field of cognitive psychology. (This straightforward separation between the two contexts has been challenged since Quine, as we have seen in the previous chapter, but we will maintain it as a first approximation for the purposes of this discussion.)

Bird (2010) provides this historical example as a nice episode illustrating how scientific knowledge accumulates by way of a continuous interplay between evidence and theory: “In the context of the debate between Millikan and Ehrenhaft over whether there is a single basic unit of electrical charge, the unique charge of an electron, Millikan’s experimental data are the evidence that helps establish the truth of the theory that electrons all have the same charge, with a value of 4.8 × 10−10 esu. In the context of assessing evidence relevant to the standard model of particle physics, these latter propositions are considered as evidence not as hypotheses.” Notice here that Bird deploys an understanding of the fact that there is no sharp difference between observation and hypothesis, with the notion that electrons have uniform charge shifting from the first to the second status depending on the theoretical context. Indeed, Bird goes on to discuss cases in which the very category of “observation” is actually quite fuzzy, as when we consider weather models, where a good chunk of the “observations” is actually constituted by the output of processing of raw data collected by automated stations and put through the filter of sophisticated computer algorithms — all without direct human intervention, and pretty removed from the everyday meaning of the term “observation.”

All of this to acknowledge that Bird’s account is anything but naive, and yet is affected by a couple of issues. One is its reliance on the already mentioned conception of knowledge as justified true belief (which in turn hinges on some version of the correspondence theory of truth, typically rejected by functionalists like Kuhn and Laudan). Another — more subtle — one is sometimes referred to as the problem of unconceived alternatives (Devitt 2011). Science is a human activity, and as such it is severely limited by the epistemic constraints inherent in being human, including the fact that at any given time scientists may simply not have thought of a good enough alternative theory (not to mention of the best theory) for any particular problem, and we have no way to know this if not a posteriori, from a historical perspective. Another way to put this is that scientists, at any particular point in time, mostly have access to “conceptual neighbors” of already available solutions or frameworks, and may be at least temporarily stuck on a local peak in the epistemic landscape. In fact, they may have simply been able to come up only with a bad lot of alternatives, and picking the best of a bad lot doesn’t really constitute knowledge, nor necessarily even mild progress.

The third account of scientific progress proposed to date is the semantic one presented by Niiniluoto (1980, 2011) and criticized by Bird (2007). Bird, somewhat grudgingly, given his propensity for epistemic theories of scientific progress, admits that semantic approaches are popular even among scientific realists (as opposed to functional accounts, usually endorsed by scientific anti-realists) because they — like the anti-realists — admit the force of the so-called pessimistic meta-induction in the history of science (Papineau 2010; Worrall 2012). This is the idea that, historically speaking, scientific theories have eventually been found to be wrong and have been replaced by new theories. [3] While the functionalist takes this as evidence that one simply cannot articulate a meaningful sense of progress in science, the semantic theorist responds that this is because — while scientific theories cannot (probably ever) be said to be “true” — they can be further from or closer to the truth.

The standard example is the transition in physics from Newtonian mechanics to Einstein’s general relativity. We know that Newtonian mechanics is wrong, in the sense of not providing an accurate account of fundamental aspects of the physical world, as it ought to if it were true. We also already know that Einstein’s theory is at the very least incomplete, and in that sense therefore it is also arguably “wrong,” since it does not mesh well with that other major theory in fundamental physics, quantum mechanics (the partial incongruence between the two is, of course, what has motivated the quest for a so-called “theory of everything,” for which string theories are potential candidates: Bailin and Love 2010; but see Smolin 2007). And yet, contra the functionalist, most scientists would want to claim that there is a sense in which Einstein’s theory is better than Newton’s, and specifically closer to the truth. How do we do that, given the problems with epistemic accounts of scientific progress based on the justified true belief concept of knowledge? One option is to turn to the semantic approach, and particularly to its core concept of “verisimilitude,” or truth-likeness: general relativity is closer to the truth in the sense of having a higher degree of verisimilitude with respect to its rival. This, of course, in turn raises the question of how to cash out the idea of verisimilitude itself.

Niiniluoto (1987), for example, developed a formal (and significantly technical) notion of truth-likeness: “closeness to the truth is explicated ‘locally’ by means of the distances of partial answers g in D(B) to the target h* in a cognitive problem B.” The first notion to unpack here is the one referring to local explication. The author rightly assumes that even if we are realists about the world (i.e., we are comfortable with the notion that there actually is a world out there with certain objective features, regardless of how much of those features can be discovered by human efforts), it will still be the case that the only way to attempt to understand such world is by deploying one or another properly suitable conceptual framework (“paradigm,” if one wishes to use Kuhn’s terminology). But, and here is the crucial point, there very likely isn’t going to be a single ideal framework capable of accounting for all phenomena in the world. This is why we have different “special” sciences (Fodor 1974), each with a certain domain of interest, tools, theories, methods, etc. That being the case, then it follows that verisimilitude is bound to be a local measure, because “more true” is quantifiable only within a given framework (or paradigm — Kuhn would have approved). This also immediately makes sense of the reference to a specific “cognitive problem” in Niiniluoto’s definition of verisimilitude given above: we are not talking about Truth in a cosmic sense, here; we are, rather, talking about the truth-likeness of notions concerning a specific and circumscribed problem, where again truth-likeness is measured relatively to currently available frameworks. [4] Niiniluoto’s approach is also sensitive to the sort of historical perspective that Kuhn first brought to the forefront of philosophy of science, and without which one simply cannot seriously engage in discussions of progress of science: “rational appraisal of theories is … historical-bound to the best conceptual systems that we so far have been able to find” (1980, 445).

A standard response from functionalists like Laudan is going to be that it is not possible to articulate a meaningful sense of “closer to the truth” unless one has an independent way of estimating where such truth lies (this is, you will recall, the same sort of objection often raised against the correspondence theory of truth: what metric allows us to measure the degree of correspondence between our theories of the world and the, by definition unknown, state of the world itself?). But it is a utopian dream to somehow be able to obtain such an independent estimate, so where does that leave the semantic approach?

Niiniluoto’s answer is that the goal is not utopian at all, and is in fact analogous to what social and natural scientists do all the time when faced with estimating the value of variables that are not directly observable (such as “fitness” in biology, or “intelligence” in the social sciences). Indeed, a family of well understood and highly functional statistical methods have been designed precisely for this purpose, methods often referred to as structural equation modeling (Kline 2011). The approach consists in constructing one or more models of the phenomenon that one wishes to analyze, in terms of sets of linear equations relating dependent and independent variables that have been measured directly. The user, however, also has the option of specifying one or more hidden (“latent”) variable, i.e., variables that are postulated to play a causal role, yet could not be measured directly. This works as long as such variables are explicitly related, in the model, to others that can be subjected to measurement and that can reasonably be used as proxies for the hidden one. As Niiniluoto concludes: “There are evidential situations e and hypotheses h such that ver(h/e) is high. In such cases, it is rational for us to claim that the unknown degree of truthlikeness … is also high, but this estimate may of course be wrong (and corrigible by further evidence)” (where ver=verisimilitude, 1980, 447).

I do not wish (and in fact I am not in a position, really) to adjudicate the ongoing debate among functionalists, epistemicists and semanticists about the nature (or existence) of scientific progress. Besides, there are several other important takes on the question of progress in science that have been proposed by philosophers over the past several decades and that I can do little more than mention in passing here, referring the interested reader to the proper technical literature. For instance, Philip Kitcher’s multidimensional account (Kitcher 1993; see update and discussion in Kaiser and Seide 2013), which stems from the author’s pragmatic naturalistic response to historiographers like Kuhn. For Kitcher scientific progress is indeed cumulative, but not in the straightforward manner proposed by Carnap and the logical positivists, since science proceeds along a number of dimensions, including the determination of natural kinds, the development of explanations, and an increasing approximation to truth (Kitcher 2012).

Another family of views on the progress of science as been termed “convergent realism,” and has been elaborated most prominently by Boyd (1973), but also Putnam (1978), and criticized by historicists and anti-realists like Laudan (1981b; see also Boyd 2007). This is the idea that scientific theories tend to be at least approximately true, with more recent theories getting closer to the truth, and that mature theories in an important sense preserve the theoretical (especially mathematical) relations of previous theories. We will examine this sort of reasoning further in a little bit, when I’ll discuss the debate between realists and antirealists in philosophy of science.

Or, finally, take Ian Hacking’s views as expressed in his Representing and Intervening (1983), which was also written in the context of that same realism-antirealism debate. Hacking’s “representing” is concerned with the variety of available accounts of scientific objectivity, but it is his “intervening” that is relevant here: in that part of the book he presents an in-depth discussion of experimental science, accompanied by a number of well developed examples (e.g., the use of microscopes in cell biology). He ends up admitting that if we limit our discussion to theoretical science, it is difficult to defend a realist position about scientific theories, which means that it is more difficult to articulate in what sense science makes progress (as opposed to being simply empirically adequate, as antirealists would maintain). But once we move to experiments, with the ability they afford us to control and manipulate systems and outcomes, it is much easier to defend the proposition that science makes actual teleonomic progress toward truth about the natural world.

What I hope the foregoing discussion has made clear is that — contra rather facile scoffing on the part of a number of publicly prominent scientists — philosophy actually has quite a bit to say about the nature of the scientific enterprise, highlighting the simple fact that even the seemingly uncontroversial idea that science makes progress is anything but. Before leaving this topic, however, we need to take a look at even more radical — and likely unpalatable to a number of scientists — philosophical ideas about the nature of science and scientific progress.

Notes

[2] However, we will see in a bit about Gettier-type objections to the standard view of what constitutes knowledge.

[3] This really only describes half of the meta-induction, and the less controversial half at that; we should note that the meta-induction concludes that all future theories will also turn out to be false.

[4] Perhaps a useful way to think about this is to realize that there is no sense in which we can say, for example, that Darwinian evolution is closer to the truth than Newtonian mechanics, since they don’t share the same framework, and are not concerned with the same cognitive problem.

References

Bailin, D. and Love, A. (2010) Supersymmetric Gauge Field Theory and String Theory. CRC Press.

Bird, A. (2007) What is scientific progress? Nous 41:64-89.

Bird, A. (2010) The epistemology of science — a bird’s eye view. Synthese DOI 10.1007/s11229-010-9740-4.

Boyd, R. (1973) Realism, underdetermination, and a causal theory of evidence. Nois 7:1-12.

Boyd, R. (2007) What Realism implies and what it does not. Dialectica 43:5-29.

Chalmers, A.F. (2013) What Is This Thing Called Science? Hackett Publishing.

Devitt, M. (2011) Are unconceived alternatives a problem for scientific realism? Journal for General Philosophy of Science 42:285-293.

Fodor, J. (1974) Special sciences (Or: the disunity of science as a working hypothesis). Synthese 28:97-115.

Hacking, I. (1983) Representing and Intervening: Introductory Topics in the Philosophy of Natural Science. Cambridge University Press.

Horgan, J. (1996) The End Of Science: Facing The Limits Of Knowledge In The Twilight Of The Scientific Age. Basic Books.

Kaiser, M.I. and Seide, A. (2013) Philip Kitcher: Pragmatic Naturalism. Walter de Gruyter.

Kitcher, P. (1993) The Advancement of Science: Science Without Legend, Objectivity Without Illusions. Oxford University Press.

Kitcher, P. (2012) Preludes to Pragmatism: Toward a Reconstruction of Philosophy. Oxford University Press.

Kline, R.B. (2011) Principles and Practice of Structural Equation Modeling. Guilford Press.

Kuhn, T. (1982) Commensurability, comparability, communicability. Philosophy of Science 2:669-688.

Kuhn, T. (2012) The Structure of Scientific Revolutions: 50th Anniversary Edition. University of Chicago Press.

Lakatos, I. (1963/64) Proofs and refutations. British Journal for the Philosophy of Science 14:1-25, 120-139, 221-243, 296-342.

Laudan, L. (1981a) A problem-solving approach to scientific progress. In: Ian Hacking (ed.), Scientific Revolutions. Oxford University Press.

Laudan, L. (1981b) A Confutation of Convergent Realism. Philosophy of Science 48:19–48.

Niiniluoto, I. (1980) Scientific progress. Synthese 45:427-462.

Niiniluoto, I. (1987) 1987, Truthlikeness. D. Reidel.

Niiniluoto, I. (2011) Scientific progress. Stanford Encyclopedia of Philosophy (accessed on 26 June 2012).

Papineau, D. (2010) Realism, Ramsey sentences and the pessimistic meta-induction. Studies in History and Philosophy of Science, Part A 41:375-385.

Pigliucci, M. and Boudry, M. (2013) Philosophy of Pseudoscience: Reconsidering the Demarcation Problem. University of Chicago Press.

Putnam, H. (1978) Meaning and the Moral Sciences. Routledge & Kegan Paul.

Rowbottom, D.P. (2011) Kuhn vs. Popper on criticism and dogmatism in science: a resolution at the group level. Studies in History and Philosophy of Science, Part A 42:117-124.

Sarton, G. (1936) The Study of the History of Science. Harvard University Press.

Smolin, L. (2007) The Trouble With Physics: The Rise of String Theory, The Fall of a Science, and What Comes Next. Mariner Books.

Worrall, J. (2012) Miracles and structural realism. The Western Ontario Series in Philosophy of Science 77:77-95.