Outline of Larry Laudan's Science and Values

Chapter 1
Chapter 2
Chapter 3
Chapter 4

I. Laudan's task (The Preface)

    A. Meaning of "values" in the Title
        1. Not about how to make science ethical or ethics scientific
        2. "Values" in title refers to cognitive, not moral, values, such as
            a) accuracy
            b) fertility
            c) predictiveness
            d) explanatory scope, etc.
        3. Cognitive values determine what methodological rules scientists ought to follow
        4. Laudan is out to defend a normative rationality:
            a) tell scientists what "theories" they ought to choose
            b) " " " "methods" " " " use
            c) " " " "values" " " " aim to achieve

    B. Relation to Kuhn
        1. since Kuhn wrote our understanding of scientific rationality has improved
        2. Kuhn's "answers" (his image of science) were never as compelling as his
        3. what we now know about scientific controversy makes them no longer plausible

II. Empiricists' View vs. "New Wave" on dynamics of Consensus/Dissensus formation (Chap 1)

    A. The Janus-like two faces of science: at any moment in history we observe that
        1. scientists agree about a lot - consensus
        2. scientists disagree over a lot- dissensus
        3. an adequate account of scientific rationality must account for both

    B. The Empiricist View's philosophical preoccupation with consensus
        1. Took cognitive agreement to be defining feature of "science"
        2. Such agreement is determined by methodological rules - the "Leibnizean ideal"
            a) methodological rules are "timeless" and define all sciences (universal)
            b) they determine what evidence will decide the question
            c) scientific disagreements are transitory
            d) explained the speed with which agreement on new theory is reached
        3. disagreed over what the rules were and/or whether we now knew them all
        4. longstanding disagreements in science could be explained by
            a) claiming sometimes scientists are irrational
            b) such disputes are really just disputes over words - different languages

    C. The empiricist view's sociological account of consensus (Mertonian sociology)
        1. Also took ability to reach rational agreement to be the defining characteristic
        2. Explained this in terms of shared "values" of scientific community
            a) empirical research alleged to discover (partially) what these values are
            b) rejection rate for submissions to science journals lower than humanities
        3. The sharing of values defines the community
        4. relative lack of shared values in social sciences means less consensus
        5. disagreements arise because other values (e.g. theological values) interfere
        6. essentially agrees with the same values philosophers found in scientific method

    D. New Wave (Kuhn and fellow travelers) Philosophical Concern with Dissensus
        1. Four lines of argument that subvert empiricist's preoccupation with consensus:
            a) ubiquity of controversy - current research shows deep controversies are
                i) long lasting common features of science
                ii) not just over the lack of evidence
                iii) not just quarrels over words
                iv) empiricist consensus cannot explain how "revolutions" get off the ground
            b) incommensurability thesis (Kuhn)
                i) historical evidence shows rival theories are committed to
                    a) different standards/values
                    b) different problems
                ii) empiricist's commitment to universal timeless method: not historically the case
            c) underdetermination of theories by data
                i) any body of data logically compatible with infinite no. of theories
                ii) Quine-Duhem thesis: any theory can be saved from refutation
                by modifying other theories in whole system of belief
            d) counternormal behavior - historical evidence indicates
                i) some most successful scientists violate alleged rules
                ii) progress often requires violating rules
        2. for these four reasons "New Wave" philosophers sought to explain dissensus
        3. because of this preoccupation New Wave cannot explain how consensus forms
        4. Kuhn specifically fails because if incommensurability is as he claims, then
            a) there can be no rationality to determine theory choice
            b) therefore actual choice is determined by Realpolitik of community
            c) at different times Kuhn says things which are inconsistent
                i) his account of paradigms implies values are not shared by rivals
                ii) other times he speaks of common timeless universal values
        5. Lakatos fails because choice of one among competing research programme's
        cannot be rationally explained
        6. Feyerabend does away with (or wants to) consensus altogether

    E. New Wave sociological concern with dissensus (Harry Collins, Andy Pickering
        1. hold theory choice entirely determined by "external" factors in society
        2. deny that "how the world really is" has anything to do with theory choice
        3. having severed reality from theory acceptance it is easy to explain disagreement
        4. in Laudan's view they cannot explain how consensus ever forms again

III: The Hierarchical Model of Justification (Chapter II)
    A. The problem of consensus is about the dynamics of convergent belief formation
        1. The Hierarchical Model is the best known answer - empiricist consensus' account
        2. In this Chapter Laudan will explicate and then criticize this model

    B. The three levels of agreement/disagreement in the Hierarchical Model (the hierarchies)
        1. The Factual level: disputes over "matters of fact"
            a) not only observational, but all claims about the world
            b) disputes are settled by appeal to methodological level
            c) at this level the rules are taken as given

    2. The Methodological Level: disputes over the rules to settle factual disputes
            a) are in effect "rules of evidence"
            b) serve as a kind of "science court" in which a "verdict" on factual disputes is resolved by appeal to "evidence"
            c) this reflects the empiricist's commitment to the "Leibnizean ideal"
            d) explains how a "staggering proportion" of disputes are in fact resolved
            e) however, occasional disputes over the methodological rules arise
    3. The Axiological Level: the level of "shared values"
            a) by appeal to shared values disputes over methodological rules are settled
            b) methodological rules are "instrumental" in character
                i) they are adopted as means to achieving certain ends
                ii) those ends are the science's "values" or "aims"
                iii) they have the forms of a "hypothetical imperative": If you desire Aim X, then follow methodological rule Y.
                iv) rules are adopted because they are believed to be optimal technique for attain cognitive goals or "utilities"
                v) choice of methodological rule is choice of most efficient means to the end of attaining our "cognitive values"
            c) disputes at axiological level are either
                i) thought to be nonexistent, or
                ii) irresolvable

    C. Factual Consensus Formation
        1. Defense of Hierarchical Model requires explaining factual consensus by rules
        2. underdetermination of theory choice by rules and evidence threatens this
            a) often rivals are "empirically equivalent" (both agree with same evidence)
            b) often existing evidence cannot dictate a choice
        3. but this problem misstates what is at issue in real science
            a) problem arises because philosopher asks which of all possible theories to choose
            b) real scientist is not confronted with all possible theories
        4. therefore the real question is which of competing rivals to prefer
        5. So understood, the question becomes how did a situation where two rivals were
            competing change into one where one of those rivals won out
        6. Rules don't "pick out the best of all possible" theories, but rather
        7. they divide potential rivals into "permissible" and "impermissible"
        8. of the actual rivals often only one is "permissible"
        9. thus even though rules underdetermine theory choice, they may dictate a preference
        10. logical non sequitur to reason from the thesis that rules underdetermine theory choice to the
        conclusion that "all theories are equal"
            a) Collins commits this when he reasons from underdetermination to the conclusion that evidence has
            nothing to do with choice
            b) Kuhn also guilty in concluding rules never determine choice
                i) Laudan distinguishes between preference among competing rivals
                ii) determining best theory; so
                iii) using Kuhn's own values he can show theory preference is rule determined
        11. Laudan proposes to amend the Hierarchical model to allow:
            a) rules may, on many occasions, determine factual preference, but
            b) sometimes they don't (even when there is agreement on the rules), and
            c) sometimes disputes over the rules arise
            d) explains the presence of both consensus and dissensus at factual level

    D. Methodological Consensus Formation
        1. Laudan understands methodological rules as instrumental (see III.B.3.b above)
        2. resolution of methodological disputes lies in appeal to axiological level
        3. the rule to choose is the one most efficient in attaining our cognitive aims
        4. but that choice is among the best of competing rivals, not all possible
        5. thus cognitive aims "underdetermine" choice of rules just as rules underdetermine choice at factual level
        6. philosopher is misguided to look for "one and only" correct scientific method
        7. historical evidence bears this out: new rules replace old when they are seen as better means of reaching a goal
        8. shared values dictate preference of competing rival rules, not best of all possible
        9. preference is underdetermined when
            a) values are shared but weighted differently
            b) values are not shared
        10. as we learn more about the world at the factual level, we may discover new techniques for
        achieving cognitive aims [this is the naturalizing move]
            a) changed understanding of sampling techniques
            b) methodological lesson of "placebo effect"
        11. Therefore, "scientific methodology" reconceived as an empirical discipline
            a) best way to achieve ends depends on the kind of world we live in
            b) a priori ("armchair") methodology is ill-conceived (the E.C.)

    E. Conclusion: In (Laudan's updated version of) the Hierarchical Model
        1. consensus forms when
            a) methodological rules dictate factual preference
            b) shared ends dictate most efficient of rival methodological rules
            c) values (aims, goals, ends, utilities) are shared
        2. dissensus emerges when
            a) shared methodological rules fail to dictate factual preference
            b) shared values fail to dictate methodological preference
            c) values are shared but weighted differently
            d) values are not shared
        3. this is far from Leibnizean ideal, but true to the "Janus face" of science

IV: Laudan's "Reticulated Model" of justification (Chapter III)

    A. Critique of Hierarchical model reveals
        1. Model is "promising" in explaining much consensus formation
        2. Needs to be stripped of "grandiose" claim to settle all debates
        3. Repeatedly breaks down badly in cases of axiological disagreement
            a) has no resources to explain how to settle such disagreements
            b) yet such disagreements arise and are repeatedly settled
        4. Therefore Hierarchical model needs to supplemented with additional machinery

    B. The covariance fallacy
        1. first form of the fallacy:
            a) assume that factual disagreements imply different axiologies
            b) committed frequently by Kuhn because his concept of paradigm implies
                i) scientists advocating rival paradigms live in different "worlds"
                ii) therefore Kuhn assumes they cannot share common values
                iii) further, he assumes that if they do share a paradigm, then
                iv) they are committed to the same goals.
        2. Second form of the fallacy (mirror image of first):
            a) assume that factual agreements imply shared axiologies
            b) committed frequently by empiricist consensus philosophers and sociologists of science
        3. Laudan's Criticism of the fallacies of covariance
            a) makes linkages between factual belief, methodological rules and goals too "tight"
            b) Thus disagreement on one implies disagreement on all...
            c) and agreement on one implies agreement on all, but...
            d) any methodological rule can achieve a variety of different goals, so...
            e) scientists can disagree over the goal of science and yet agree on rules.
            f) This explains how disagreements over realist and instrumentalist goals
            g) are compatible with agreement over methodological rules.
            h) Therefore, axiological consensus is neither
                i) necessary for factual consensus, nor
                ii) sufficient for factual consensus because
                iii) axiological goals underdetermine methodological rules

    C. Temptation to commit covariance fallacy explains
        1. tendency to assume there is more axiological consensus than in fact is the case
        2. tendency to assume that axiological disagreements are irresolvable rationally
            a) Popper and Reichenbach guilty of this (on the RV side)
                i) with Popper methodological rules are conventions
                ii) goal of ever truer theories is matter of "taste" (convention)
                iii) only demand is that values be "internally" consistent
            b) also Kuhn on the "New Wave" side
                i) add Kuhn's claim that rival paradigms have rival values to
                ii) RV view that values are conventions, and we get the conclusion
                iii) paradigm shifts are necessarily irrational
                iv) therefore history of science is part of the history of fashion
        3. history of science does not reveal such irresolvability rationally because
            a) whole factions are not arbitrarily "banished" from science
            b) loosers often come to enthusiastically accept values of winners

    D. Laudan's analysis of the rationality of axiological debate
        1. Not always possible to settle rationally axiological disagreements, but
            a) in a "sufficiently large range of cases" it is possible, so
            b) the general view it is always impossible is highly misleading
        2. Goals can be rationally criticized as "utopian"
            a) "rationality" implies belief that the goals you seek are attainable
            b) belief that a goal is attainable depends on factual beliefs about the world
            c) as factual beliefs change, a goal thought attainable may cease to be
            d) three kinds of "utopian" goals
                i) demonstrable utopianism - goal logically unattainable
                ii) semantic " - goal too ill defined to tell if attained
                iii) epistemological " - no way to tell if goal has been attained
        3. Goals can be rationally criticized as failing to accord with implicit values
            a) goal explicitly avowed aren't necessarily those implicit in practice
            b) a scientist can be unaware of his implicit goals - false consciousness
            c) can be criticized on grounds of practice what you preach
                i) bring implicit goals to accord with explicitly professed goals
                ii) or change explicit goals to agree with implicit one in practice
            d) if you can't be with the one you love, love the one you're with.
            e) so if you can't achieve the goals you espouse, espouse the goals you can
        4. Extended example of axiological change
            a) status of "hypothetical method" in 17th century physics
                i) in Newton's time it seemed to shun hypothesizing unobservables
                ii) 100 yrs later such methods were having success
                iii) explicit professed goals changed to accord with practice
            b) Descartes' goal that causes must be more clearly known that effects
                i) a "lackluster performance" in achieving such goals
                ii) came to be abandoned because proved unattainable

    E. Laudan's "Reticulated Model"
        1. Theories, Methods, and Aims (goals, ends, values) not in a hierarchy
        2. are all interlinked such that
            a) changes in one can (but may not) affect others
            b) do not necessarily change together (do not "co-vary")
            c) connections are much "looser" than in Hierarchical model
        3. Theories and methods affect each other because
            a) methodological rules select which theories we accept
            b) theories we accept about world constrain what methods are believed to be instrumentally effective
        4. Methods and Aims affect each other because
            a) aims justify which methods we employ
            b) methodological rules exhibit realizability (or not) of aims
        5. Theories and Aims affect each other because the two must "harmonize"
        6. Laudan's theory is "thoroughly Heraclitean" - change at all levels
        7. No one aim (e.g. "truth") is "the right" aim of science
        8. Problem of progress answered
            a) progress is always relative to some end
            b) as ends change what constitutes progress changes
            c) what we call progress is progress towards our goals
                i) earlier scientists may not have shared these goals
                ii) still their work could be progress towards them

V. Laudan's Critique of the "Holist Picture of Scientific Change" (Chapter IV)

    A. approach
        1. use Kuhn as "stalking horse" to analyze problem of scientific change
            a) will show the superiority of Laudan's reticulational model
            b) though Kuhn has cleared up ambiguities and appears more conservative
                i) he still cannot explain the rationality of scientific change
                ii) he has also failed to describe the historical process correctly
        2. others also guilty of same mistakes, including Lakatos and younger Laudan

    B. Kuhn's commitment to paradigms as "units" of scientific change
        1. each paradigm contains its "own"
            a) ontology (world-view) (i.e. factual consensus)
            b) methodology, (i.e. consensus on rules)
            c) axiology (i.e. consensus on values)
        2. they form an "inextricable mix"
        3. Kuhn excludes possibility of internal change within the paradigm
        4. So, when one changes, all simultaneously change (Gestalt switch)
        5. Kuhn's notion of paradigm "internalizes" the hierarchical model so well that
            a) whenever hierarchical model breaks down, so does Kuhn's
            b) makes paradigm change an abrupt rupture of the scientific life
            c) change is irrational "conversion" experience
        6. Kuhn gives paradigms an "integral" (holist) and "static" character; thus
        7. Kuhn ignores the possibility of rational change within the paradigm
        8. Thus he makes paradigm change essentially change of fashion, and
        9. such changes are the result of "external" factors not rational ones
        10 Kuhn guilty of relativism, subjectivism, and irrationalism
        11. These are "high sins" on the philosophers list because...
        12. for two millennia philosophers have been at pains to show such disputes can be resolved by reason

    C. Loosening up the fit
        1. Laudan's two proposals:
            a) replace hierarchical model with reticulational model
            b) simply drop Kuhn's insistence on integral (holist) character of paradigm
        2. On reticulational model ontology, methodology, and axiology interact loosely
        3. A scientist may have good reason to
            a) change factual beliefs (ontology) without methodology or axiology data changes so rules
                and aims dictate new factual beliefs
            b) change methodological commitments without ontology or axiology discover a new method
                justifying the same theories which is more efficient means of reaching old aim
            c) change axiology without ontology or methodology start professing what you been practicing
        4. In all above cases there is "enough common ground" to make rational choices

    D. Why has this sort of change not been seen by Kuhn et al.?
        1. tendency to "telescope" history
            a) from vantage of a century later 20-30 years appears as a single stage
            b) thus a gradual process of change can be made to appear abrupt
        2. Laudan's simplified possible history of gradual change (his "tall tale"):
            a) existing methodology and aims dictate change in factual belief
            b) new theory (factual beliefs) replaces old theory, i.e. new ontology
            c) new factual beliefs lead to discovery of more efficient method
            d) new methodological rule replaces old rule, i.e. new methodology
            e) old aims, in light of new factual beliefs and new rules, now seen as
                i) unattainable (utopian)
                ii) or inconsistent with practice (out of harmony)
            f) new aims already implicit in practice replace old, i.e. new axiology
        3. So even though such a series of steps is rational
            a) if we look only at "before" and "after" the single period, it seems
            b) all three levels have abruptly changed, and so
            c) "paradigm shifts" (as Kuhn describes them) appear irrational
        4. Kuhn guilty of the "fallacy of telescoping history" (which every historian is taught to avoid)

    E. Is this "gradualist" view of change really the case?
        1. of course real history isn't as simple as this "fairy tale" (D.2 above)
        2. above case considers a single "tradition" or paradigm
        3. more challenging case is more like real history of multi-paradigm changes
        4. because of underdetermination of theories by rules and rules by aims
            a) a rival's theories may come out better than yours even on your standards
            b) a rival's methods may be more efficient even on your aims
        5. so advocates of one paradigm may take over a rival's theories, methods, or vaims, while leaving others unchanged
        6. so multiparadigm (tradition) changes reduce to many unitraditional changes
        7. the core question now becomes, does the historical record support
            a) the Laudanian, reticulational, gradualist rational picture, or
            b) the Kuhnian, hierarchical, holist irrationalist picture
        8. Laudan offers as evidence
            a) history of "hypothetical method" (unobservables)
                i) an example of a major methodological shift
                ii) no accompanying axiological or world-view shift
                iii) this shows methodology can change without others changing
            b) history of "certainty" as a goal replaced by "probability"
                i) a major axiological shift in the goal of science
                ii) no parallel shift in world-view or methodology
                iii) this shows axiology can change without the others changing
        9. Laudan's picture of change is "piecemeal" rather than "holist"
            a) no evidence of striking co-variance of change among three levels
            b) three levels do not come as "an inseparable package"
            c) not accepted "on a take-it-or-leave-it basis"
        10. the "great revolutions" of science do indeed involve changes in all three
            a) this led young Laudan and Kuhn to see this change as all three connected
            b) now he sees these change as sequential, not simultaneous
        11. claims for irrationality of paradigm choice systematically ignore
            a) piecemeal character of scientific change
            b) "adjudicatory mechanisms" which make such changes rational

    F. Kuhn's four arguments for underdetermination:
        1. Kuhn's attack on rationality relies on his claim that methodology and aims
            a) always underdetermine theory choice
            b) can't even decide "locally" which of competing rivals to prefer
        2. ambiguity of shared standards argument
            a) aims so ambiguous that scientists can interpret in variety of ways
            b) so disagreements arise as to how to apply such standards
            c) so any theory can be preferred, so scientists cannot agree
            d) Laudan's reply:
                i) this may sometimes be true
                ii) it is hardly always true
                iii) consistency is an example of a non-ambiguous standard
                iv) controlled testability and fertility other non-ambiguous standards
        3. collective inconsistency of rules argument
            a) collectively rules and standards often conflict with each other
            b) different scientists will weight different conflicting standards differently
            c) so they cannot agree which theory to prefer
            d) Laudan's reply
                i) grants that goals often conflict, indeed emphasizes it
                ii) Kuhn has not shown that all methodologies have conflicting goals
                iii) Kuhn has globally extrapolated from a small handful of cases
        4. shifting standards argument
            a) Kuhn holds advocates of rival paradigms must have different standards
            b) so each is to be preferred by its own standards
            c) Laudan's reply:
                i) commits the covariance fallacy of assuming difference in factual belief demands different standards
                ii) historical record denies this holist character
        5. problem-weighting argument
            a) advocates of rival paradigms will disagree over which problems are most the important to solve
            b) since different paradigms will be successful in solving different problems
            c) no rational decision which to prefer
            d) Laudan's reply:
                i) such differences do surely exist but they do not
                ii) do not invariably imply they cannot be rationally resolved
                iii) Kuhn confuses problem's importance as evidence for theory with
                iv) social or external importance of solving problem
                v) what matters is a problem's epistemic importance, i.e. its importance in making the theory beliefworthy
                vi) while scientists differ in motivational importance of problems
                vii) and so this may be a matter of taste and arbitrary
                viii) the scientist is not "free to choose" which problem solutions do provide support for a theory
                ix) how much solving a problem supports a paradigm is not simply
                x) how keenly scientists want to solve the problem
                xi) how much support a problems has is a matter of objective methodology not subjective interest
                xii) how well which of rival theories pass evidential tests not ameasure of how important the
                    scientist finds the problem

        6. Summary: Though what recruits a scientist to a paradigm is subjective and pragmatic, how rational we are
        to prefer one paradigm over another rests on "principles of empirical or evidential support which are neither
        paradigm specific, hopelessly vague, nor individually idiosyncratic [and which are]...sometimes sufficient to
        guide our preferences unambiguously.