What are Scientific Revolutions? (1987)
Kuhn divides scientific development into normal and revolutionary, and focuses on the revolutionary, or noncumulative mode. He describes three characteristics which are shared by all revolutionary changes (14 top) and uses three extended examples to illustrate these characteristics.
He describes the gestalt switch that enabled him to understand Aristotelian physics, a system which rests on a very different understanding of "change" and of "inherent qualities." The second example is based on Voltaic cells. As it was originally conceived, a battery unit consisted of two pieces of metal in contact with each other; as it was reconceived later, a battery was a liquid separating two metals. The third example is drawn from Max Planck's work on the black-body problem. Planck borrowed the mathematical solution that Boltzmann used for understanding the behavior of a gas, and applied that solution to explaining why the color of a heated body changes with temperature. Six years later his mathematics, and his conception of the placement of what he called "resonaters," was re-figured to include a different model, with discontinuous jumps of energy, and a new term, "oscillators," to replace resonaters.
The first characteristic is the gestalt switch, the blinding flash, in the mind of the researcher. "Revolutionary changes are holistic." (28) The second characteristic is the ability to re-conceive constituent elements of a problem by first adding a new word for a newly discovered property, and second, changing the way that all other established vocabulary relates both to the new word, and also to nature itself. The third characteristic is a change in governing metaphors. A metaphor contains the taxonomy and salient features of a phenomenon, expressing by comparison and contrast what a thing can and cannot do.
Commensurability, Comparability, Communicability (1982)
Kuhn defends and explains his understanding of commensurability, comparability, and communicability.
Incommensurability means, literally, "no common language," but Kuhn means something more limited: that "there is no language, neutral or otherwise, into which both theories, conceived of as sets of sentences, can be translated without residue or loss." (36) This is local incommensurability. Different cultures, different historical periods, possess different lexicons, or structured vocabularies. These lexicons may overlap, and may enrich each other.
He answers his critics, Davidson, Kitcher, and Putnam, who, according to Kuhn, mistakenly dismiss incommensurability because they sketch techniques of interpretation, but mislabel these techniques translation. (37) Translation is done by someone who knows two languages. The translator finds a term, or set of terms, into which the phrase can be rendered. Any foreign language can be understood by a non-native , but only because the non-native has become bilingual. Interpretation, by contrast, is "practiced by historians and anthropologists." (38) Interpreters, commanding only a first language, make a hypotheses of meaning based on observable action. If necessary the interpreter borrows the new word in the unfamiliar language to become a word in the home language. Interpretation adds new concepts, altering the relationship between words and creating new systems for relating words to each other. In this case, the two languages were, and remain, incommensurable in their original, un-interpreted state. A language can only be enriched by interpretation as far as its governing ontology will permit. (54) Kuhn says that W.V.O. Quine's examples in Word and Object (MIT Press 1960), conflate translation and interpretation. The historian delving into old science texts, and encountering alien systems of thought, alien understandings of how the world works, or ought to work, serves as a translator.
Working out the details of his argument, and addressing Kitcher, Kuhn uses the example of phlogiston. He adds a second example using the Newtonian concepts of mass and force -- suggesting that in order for either term to make sense, they must be learned together. To explain an older, and out-dated system, the historian must simultaneously translate those terms that have modern referents, and interpret those terms that have no modern referents.
In a point which I couldn't follow, Kuhn invokes Ramsay sentences (46). He pauses to address Quine's "translation manual," (parallel lists of words and phrases). Context specifiers are the relationships, known to members within a culture, that specify which words are regularly linked to other words in the language . (48) When a word belongs to a cluster of interrelated terms, then changing the meaning of one changes the meaning of all; and hence there cannot be a perfect translation from one system to another. However, different speakers from the same culture, who have different criteria for words and their referents, still generally pick the same terms and referents together because -- Kuhn argues -- some sort of local cultural "holism" must be a feature of language (52). Different languages impose different structures on the world. (52)
A new theory of meaning will include "tropes" and "taxonomy," or a knowledge of shared categories and shared relationships between them. A sign that many in the community have changed paradigms could be the holistic language change that accompanies a paradigm shift.
Possible Worlds in the History of Science (1986)
"To understand some body of past scientific belief, the historian must acquire a lexicon that here and there differs systematically from the one current in his own day." (58) Only by using the older lexicon can a historian speak for that older science accurately. Translation will not work. The essay has four sections: in section one, the idea of the lexicon is explicated; in section two the possible-world semantics debate is sketched; in section three Newtonian mechanics illustrates Kuhnian lexicon theory; and in section four the possible-world debate is applied -- in a restricted fashion -- to scientific development. (58-59)
Section One: Describing and defining the lexicon.
Kuhn argues that the past is a possible world, and that each world has its own lexicon. "Different lexicons -- those of different cultures or different historical periods, for example -- give access to different sets of possible worlds, largely but never entirely overlapping." (61) He also argues that an important aspect of the lexicon is the understood or assumed relationship between words. "Knowing what a word means is knowing how to use it for communication with the other members of the language community within which it is current. But that ability does not imply that one knows something that attaches to the word by itself, its meaning, say, or its semantic markers. Words do not, with occasional exceptions, have meaning individually, but only
through their association with other words within a semantic field. If the use of an individual term changes, then the use of the terms associated with it normally changes as well....Two people may use a set of interrelated terms in the same way, but employ different sets (in principle, totally disjunct sets) of field coordinates in doing so. " (62-63)
Section Two: Possible Worlds
We can imagine possible worlds that are similar or different to our own, and in these worlds, Kuhn argues for three kinds of "truth." "[T]he concept of the possible world offers a route both to a logic of modal statements and to an intensional semantics for logic and for natural languages. Necessarily true statements, for example, are true in all possible worlds; possibly true
statements are true in some; and a true counterfactual is a statement true in some worlds but not in that of the person who made it." (63-64)
Section Three: An Example Using Newton
Kuhn describes several ways of understanding Newtonian force, mass and weight. Antecedent vocabulary must be in place, and the three terms must be learned together. Kuhn shows that the rules for the relationship between these three terms can be created from various "angles," using different kinds of instrumentation (spring balance, pan balance). A significant re-working of the relationship between force, mass and weight, however, will require a different lexicon, or perhaps an enrichment of the old one with terms from a new lexicon.
Section Four: Limits of Possible Worlds for Science
"Faced with untranslatable statements, the historian becomes bilingual, first learning the lexicon required to frame the problematic statements, and then, if it seems relevant, comparing the whole older system (a lexicon plus a science developed with it) to the system in current use." (77) A statement can only be known as true if it is placed within a lexicon of meaning. If, as standard forms of realism suppose, a statement's being true or false depends simply on whether or not it corresponds to the real world -- independent of time, language, and culture --then the world itself must be somehow lexicon dependent" (77)
(Surely Kuhn was aware that he had invoked the phrase "in the beginning was the Word...?")
Kripke and Putnam suggested that words only relate to each other, and not to things outside themselves. Kuhn contends that a lexicon presumes relations and at least theories about the outside world: for instance, natural kinds -- things with one essence, and correspondingly short names (e.g. water) -- and essential properties. He argues that gold -- atomic number 79 -- could not be gold if it were blue, because in atomic theory the number "79" precludes the characteristic "blue." (84) For the historian, possible worlds are past worlds, and a new lexicon must be learned. "[T]here is no basis for talk of science's gradual elimination of all worlds except the single real one." (86)
Postscript: Kuhn admits an earlier error. Yes, an individual, but not a community, can undergo a "gestalt switch" Community beliefs can better be judged by community lexicons.
The Road Since Structure (1990)
This article offers a brief and "dogmatic" sketch of the main themes of a book that Kuhn was working on (unfinished when he died). In the intended book he wished to address the problems left over from the Structure of Scientific Revolutions and to study the problems accompanying the (1960s?) transition to the soft, or historical philosophy of science. (A transition which included, among others, contributions from Kuhn, Paul Feyerabend, Russ Hanson, Mary Hess, Michael Polyani, Stephen Toulmin.) Incommensurability is the key problem, the "essential component of any historical, developmental or evolutionary view of scientific knowledge." (91). Rather than a "threat to rational evaluation of truth claims," it "restore[s] some badly needed bite to the whole notion of cognitive evaluation." (91) He suggests that his ideas defend against the "excesses of postmodernmist movements like the strong program." (91)
In this article, then, he will answer these questions: what is incommensurability? What is its relationship to relativism, truth, and realism?
Earlier Kuhn described incommensurability as difficulties in learning the language of the past, or of another culture or field of study. But Kuhn now feels that "language learning" is too broad an analogy. Incommensurability has to do with the meanings of a particular class of words, the taxonomic terms, "kind" terms (including natural, artifactual, social and others). Examples might include water, sheep, dirt, fluorescent light, Hindu. In English, they are the words that take an indefinite article, the count nouns, the mass nouns. A further distinguishing point: they do not overlap in their referents, unless the two words are related as species to genus (So there are no dogs that are also cats, but compare the terms ungulate, sheep). (92). If members of a language community do encounter a dog that is also a cat, they must re-create taxonomy to incorporate the animal (Kuhn uses the platypus). Incommensurability means that terms cannot be translated. Taxonomies, or words in a particular language and how those words are related to each other in a conceptual scheme, limit the kinds of things that can be said, conceived, and believed. (93-94).
Speaking of relativism and rationality, it is no longer possible or necessary to ask whether scientific truth corresponds to an external mind-independent world. "Rather, what's to be evaluated is the desirability of a particular change-of-belief, a change which would alter the existing body of knowledge claims so as to incorporate, with minimum disruption, the new claim as well....Justification does not aim at a goal external to the historical situation but simply, in that situation, at improving the tools available for the job at hand" (96) [Wow -- would you trust this view of truth in the hands of anyone less good than God?]
Kuhn sees fruitful analogies in evolutionary theory. Every new scientific revolution splits off a new species from the parent genus. After a while, no communication with the parent genus (discourse) is possible, nor is cross-breeding with other species (intercourse). Specialization and narrowing, new lexicons, and the limits that go with lexicons, are the isolating inevitable results, but these limits are "required for the development of knowledge." (99)
Truth or falsehood are determined by the limits of the lexicon, but all communities have such lexicons, and all set up a true/false game. True/false communication breakdowns are "symptoms of the speciation-like process through which new disciplines emerge...And it's the need to maintain discourse, to keep the game of declarative sentences going, that forces these divisions and the fragmentation of knowledge that results." (101) Advancement of knowledge -- which Kuhn evidently sees as "good" -- is the "game of declarative sentences."
Even if it is not possible to know about a mind-independent world, Kuhn makes a "Kantian" move of faith: the world is not mind-created because apparently it was here before we got here and we inherit the things that others before us have thought and done. The real world is the real "environment, the stage, for all individual and social life." We affect the world; it affects us; creatures and niches evolve together. (102). The community holds primacy over the individual in so far as the individual must share a taxonomy with others in the community. However, "[u]nderlying all these processes of differentiation and change, there must, of course, be something permanent, fixed and stable....ineffable, undescribable, undiscussible. Located outside space and time this Kantian source of stability is the whole from which have been fabricated both creatures and their niches, both the 'internal' and the 'external' worlds." (104)
The Trouble with the Historical Philosophy of Science (1991)
Kuhn and his generation were taught the positivist view of science: that as history progressed, scientific facts and theories corresponded more and more closely with reality. Kuhn and his colleagues were dissatisfied with the so-called facts, and more so with the shockingly partisan method by which these facts had been established as facts. But there was trouble building a new edifice:
"The trouble with the historical philosophy of science has been, I've suggested, that by basing itself upon observations of the historical record it has undermined the pillars on which the authority of scientific knowledge was formerly thought to rest without supplying anything to replace them. The most central of the pillars I have in mind were two: first, that facts are prior to and independent of the beliefs for which the are said to supply the evidence, and second, that what emerges from the practice of science are truths, probable truths, or approximations to the truth about a mind- and culture-independent external world." (118)
Amongst those challenging the way beliefs about Nature are negotiated and concluded was a school antithetical to Kuhn's -- here I will say beliefs -- because he calls the school's conclusions both "inescapable" and "unacceptable." (111)
"the most extreme form of the movement, called by its proponents 'the strong program,' has been widely understood as claiming that power and interest are all there are. Nature itself, whatever that may be, has seemed to have no part in the development of beliefs about it. Talk of evidence, of the rationality of claims drawn from it, and of the truth of probability of those claims has been seen as simply the rhetoric behind which the victorious party cloaks its power." (110)
Kuhn proposes that the problem to understand is not the rationality of belief, but rather why a group elects to change its beliefs -- in other words, "understanding....the rationality of incremental change of belief." (112) He offers three reasons to focus on incremental change, on comparison between belief systems, and not on comparison between the Truth (capital T) of Nature and the old system. First, one needs only to understand what the old beliefs were, and why the group is changing some of those beliefs; second, one needs only to identify (afterwards) the key step, or corner, in the context of other steps or turns already taken; third, one can apply just the criteria that positivists would use: "which of the two bodies is more accurate, displays fewer inconsistencies, has a wider range of applications, or achieves these goals with the simpler machinery..."(114)
He asserts three views: "the Archimedean platform outside of history, outside of time and space, is gone beyond recall....comparative evaluation is all there is." Scientific development is like Darwinian evolution, "pushed from behind rather than pulled toward some fixed goal...." (curious, that view. How is evolution "pushed"? ). And truth is not a mind-external truth. [In reaching to express his ideas about Nature as a product of belief, Kuhn says something perhaps unintentionally amusing: "All past beliefs about nature have sooner or later turned out to be false." (115) A sentence like that makes better sense from an Archimedean platform, but oh well, we've read far enough to get the idea.] He continues, Truth, or knowledge, has advanced through field specialization (or speciation as described above). Knowledge evolution happens as new groups form, and these groups are governed by familiar human motives. Respect for power and authority are among those motives, but so, too, are the desires to be rational, to find solutions that are more accurate, broader, simpler, etc. (118) Scientists produce and evaluate change of belief, and they use criteria that correspond to rationality in their own belief-system, but not to a an external real world. Science is not a monolith, but a sprawling, branching, growing pluralistic structure.
In a "coda" Kuhn notes again his views on the analogy between revolutions and speciation; his understanding of incommensurability; and his idea that niches and niche inhabitants grow together.
Reflections on My Critics (1970)
Kuhn responds to a gathering of seven of his critics, Watkins, Toulmin, Williams, Popper, Masterman, Lakatos, and Feyerabend. Their papers and his answer were published between 1965-1969. As Kuhn sees it, there are three main charges to his account: a misunderstanding of what the scientific method should be; a misunderstanding of what normal science is, and an acceptance of "mob rule" rather than rationality. I am choosing to simplify: I am not going to recreate his "seriatim" responses, or explain (or criticize) how Kuhn groups these seven writers' views under each of the three charges. Rather, I am simply going to state Kuhn's positions on each charge.
First, theory choice depends on social context, not truth. "Whatever scientific progress may be, we must account for it by examining the nature of the scientific group, discovering what it values, what it tolerates, and what it disdains." (131) Rather than debating what criteria scientists ought to use in order to justify theory choice, it is more useful to ask simply what criteria they do use. Kuhn argues that he and his critics alike use sociology, psychology, history and context, not scientific theory, to understand choices that scientists make.
Second, normative science is not a pernicious indoctrination nor the
province of third-rate minds.
It is necessary to do work in between revolutions. In the arts, every practitioner aims at a revolution, at a new understanding of the human condition (But Kuhn suggests that some arts, such as economics and psychology, could resemble the sciences by reaching a normative state. 139). There are four criteria which normative science must satisfy: (1)"for some range of natural phenomena, concrete predictions must emerge..."(2) "for some interesting subclass of phenomena, whatever passes for predictive success must consistently be achieved." (3)"predictive techniques must have roots in a theory which, however, metaphysical, simultaneously justifies them, explains their limited success, and suggests means for their improvement in both precision and scope." (4) the improvement of predictive technique must be a challenging task, demanding on occasions the very highest measure of talent and devotion." (139) Kuhn labels himself a Popperian in this area. When do we know that a revolution has taken place? Or, in Kuhn's words, "Can we distinguish mere articulations and extensions of shared belief from changes which involve reconstruction?" (145) Answering that question, Kuhn says, can only be done by asking the all-important question, "revolution for whom?" Defining the group of scientists and their shared normative beliefs helps to answer the question "revolution?"
Third, Kuhn repudiates the view that in theory choice, "might makes right." He says that he has been misread if anyone has understood him to advance such views. Rather, he believes that in a disagreement, neither group "has access to an argument which resembles proof in a logic or formal mathematics." (156) Scientists use mathematics and logic, but they are the values used in making choices rather than the rules of choice themselves. (157) Kuhn is also not a relativist in the sense that "one scientific theory is as good as another," and he suggests that accuracy, scope, simplicity, fruitfulness are the criteria to be used.
Finally, Kuhn reiterates his views on incommensurability (outlined in greater detail above). He also gives a more thorough gloss on Quine's "gavagai/bavagai" example. He explains that paradigms -- learned similarity of relationships -- can function in absence of theory (criteria). Paradigms should be understood as community structures, or disciplinary matrices, and that students learn these matrices by ostension, by being given problem exemplars to work out. And returning to incommensurability, paradigms are a "language conditioned," or "language- correlated way of seeing the world." (171)
Theory Change as Structure Change: Comments on the Sneed Formalism (1976)
Sneed formalism makes it possible to explore with precision theory formation and change. Sneed discusses three kinds of models, "potential partial models" (Mpps), "partial models" (Mps), and "models" (Ms). As part of science, students are taught how to move from potential partial models to partial models. Also, "great" scientists discover the ways that a potential partial model can be solidified into a model, or related to an existing model.
Within a model (theory), Sneed posits a set of applications -- there must be at least a pair -- which must be tested and guided by constraints -- again, always applied to two applications. "Adequate specification of a theory must include specification of some set of exemplary applications." (179) When applications are tied together, they are capable of specifying two things: 1) "the manner in which theoretical concepts or terms must be applied;" and 2) "some empirical content of the theory itself." (180) Sneed's theory allows us to avoid the question of concept primacy (two concepts always must be present). Kuhn makes a point about the new form Ramsey sentences in Sneed's formalism...which was Greek to me. Another correspondence between Kuhn and Sneed's model is that the theory core is "immune to direct falsification." (182).
Sneed describes theory change as an expanded theory core. "A [new] function or concept becomes "theoretical with respect to a given application if constraints are required to introduce it there." (185) A complete core must be "rich enough to permit the evaluation of theoretical functions." (187) Kuhn offers a counterfactual example using Newton's law of gravity, explaining when that law would be in the theoretical core, and when it would be outside of the core. A theory of the core allows Kuhn to look at replacement and expansion more carefully. Roughly, one member of a pair of applications in a theory must be replaced by another, and quoting Sneed, Kuhn says "the new theory must be such that the old theory reduces to (a special case of) the new theory." (188) Kuhn concludes with a reiteration of his reference and translation metaphors, and an application of those ideas to models, partial models and potential partial models. If each model has a corresponding term, then incommensurability disappears -- but Kuhn takes it that such correspondence cannot exist. Again, he uses a Newtonian example (a bit beyond me) but suggests that "membership in a class of intended applications I cannot be given extensionally, by a list, because theoretical functions would then be eliminable, and theories could not grow by acquiring new applications. In addition, [Sneed] doubts that membership in I is governed by anything quite like a set of necessary and sufficient conditions" (193) -- the very features of a paradigm.
Metaphor in Science (1979)
In this paper Kuhn is responding to "What is 'Metaphor' a Metaphor For?"
presented by Richard Boyd in 1977. Kuhn believes that metaphors will
help to explain what happens when new terms are introduced into the vocabulary
of science. Causal theory says that words refer to and define things,
for example the name "Sir Walter Scott." But how to decide what is
the central definition of a thing, and what word is consequently most central
to that thing? Causal theory suggests that words are "dubbed" as
tags or labels, attached as a product of history, not immanence.
Kuhn differs with Boyd over the term "epistemic access," which Boyd uses
to refer to words that cover artifactual natural kinds (a galvanometer,
for example). Ostension, learning and applying names to new situations
with like kinds, can cause difficulties as well. Metaphors, Kuhn
suggests, are "essentially a higher-level version of the process by which
ostension enters into the establishment of reference for natural-kind terms....Metaphors
play an essential role in establishing links between scientific language
and the world." (201-203) Metaphors allow a way to compare theories
that are, by terms, incommensurable. And they suggest that there
is no sufficient reason within nature itself to choose one metaphor, or
one theory choice, over the other. Kuhn disagrees with Boyd, who suggests
that metaphors "cut into the natural joints" in nature. As close to reality
as Kuhn will sail sounds like this: "Conceived as a set of instruments
for solving technical puzzles in selected areas, science clearly gain in
precision and scope with the passage of time." (206) The world is a real
as we agree to think it is -- but there are not natural joints.
Rationality and Theory Choice (1983)
Kuhn presented a brief talk at a symposium given in honor of G.C. Hempel. Hempel is one of Kuhn's friends and colleagues, and Kuhn is here concerned with the justification of theory choice. Scientific terms must be learned in clusters. None of the terms are "independently available for use in a definition of the other" (212) but they can be tested. The very names of the disciplines gathered at a university must themselves be learned in a cluster, defined against one another. Kuhn asserts that the "usual norms for theory choice are justified" because rationality and justification are themselves interdefined terms. He admits that the logic of his structure is a bit shaky, and calls "not for a justification of learning from experience, but for an explanation of the viability of the whole language game that involves induction and underpins the form of life that we live." (215)
The Natural Human Sciences
Kuhn responds briefly to a paper by Charles Taylor "Interpretation and the Sciences of Man." Kuhn tells of reading Ernst Cassirer and Max Weber, both of whom described regular revolutions in thinking in the social sciences were a regular feature of those fields, but subjoined (and concluded) that "Die Naturwissenschaften...sind ganz anders."(216) This is the point that Kuhn wishes to challenge. Kuhn argues that the natural and human sciences bear more resemblance to each other than Taylor credits.
Kuhn and Taylor agree that a person can operate within a paradigm without
being able to name or describe the "necessary and sufficient conditions,"
or constraints, of that paradigm. However, Taylor suggests that the sciences
cannot "revolve" because Nature does not revolve. For example, the heavens
are the heavens to every culture and time. That's ripe stuff for
Kuhn, who certainly believes that what planets are and what planets were
has changed considerably.
Kuhn thinks there is a line, but that the line between human and natural sciences, but that this line should be drawn between sciences with a "hermeneutic enterprise," (the human sciences who develop always newer and deeper interpretations) and those who simply operate within the hermeneutic without needed to challenge it. He wonders, whether "over time, an increasing number of specialties will find paradigms that can support normal, puzzle-solving research?" (222)
Kuhn addresses an array of nine papers all offered at a symposium in his honor at MIT in 1990. Necessarily (or naturally) he is less inclined to be sharp with these than the seven he surveyed in Reflections on my Critics. He alludes to the circumstances that brought him to Princeton, including his warm reception by G. C. Hempel, and the intellectual problems of theory choice that they both have explored. Kuhn defines incommensurability (see above) and states that he has had "rapid breakthroughs" (228) in the latter half of the 1980s. Among those breakthroughs were a new understanding of natural kinds and a theory of lexicons. He further emphasizes the difference between language learning and translation and compares his lexicon with Jed Buchwald's conception of an unarticulated core. He turns to the shared scientific culture and comments on Norton Wise's treatment of individuals as groups, and vice-versa, as being a problem he himself mis-managed in Structure. Kuhn asserts that no lexicon can ever be mind-independent, just as it is always world-independent. Antecendently available terms (words that already exist in a lexicon) form the only basis of reality that cultures and groups can know. Finally, he argues, as above, for the obvious "rationality of the standard list of criteria for evaluating scientific belief." (251)
[Kuhn refers to synchronic and diachronic usage of terms (p.235) -- diachronic being "the resort to modern terminology;" synchronic I suppose the use of a term within a lexicon or paradigm -- not sure I understand this passage -- would like to check with you!]
A Discussion with Thomas Kuhn (1995)
A sixty page transcript of a three-day interview, this rough article
gives some of Kuhn's early family life and education. It covers his
studies at Harvard, his war-work, his graduate studies, his move to Berkley,
to Princeton and to MIT. He mentions those who served as mentors,
(Conant, Hempel) or who were sympathetic and encouraging, and also those
who were less friendly. This article was not hard reading, and useful for
anyone who wants to understand Kuhn's own impression of the context in
which The Structure of Scientific Revolutions was produced. One gets the
impression of a man who didn't make friends easily, who took offense perhaps
a little too quickly, and who was hampered by knowing what he wanted to
do but not being able to articulate those purposes easily. One senses
also a man who was hungry for intellectual peers but afraid of them when
he found them.