strong. john stuart mill, john herschel, and the probability of causes (article)

Upload: blaise09

Post on 03-Jun-2018

213 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    1/12

    John Stuart Mill, John Herschel, and the 'Probability of Causes'Author(s): John V. Strong

    Reviewed work(s):Source: PSA: Proceedings of the Biennial Meeting of the Philosophy of Science Association,Vol. 1978, Volume One: Contributed Papers (1978), pp. 31-41Published by: The University of Chicago Presson behalf of the Philosophy of Science AssociationStable URL: http://www.jstor.org/stable/192623.

    Accessed: 11/02/2013 19:34

    Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at.http://www.jstor.org/page/info/about/policies/terms.jsp

    .

    JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

    of scholarship. For more information about JSTOR, please contact [email protected].

    .

    The University of Chicago Pressand Philosophy of Science Associationare collaborating with JSTOR to

    digitize, preserve and extend access to PSA: Proceedings of the Biennial Meeting of the Philosophy of ScienceAssociation.

    http://www.jstor.org

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PM

    http://www.jstor.org/action/showPublisher?publisherCode=ucpresshttp://www.jstor.org/action/showPublisher?publisherCode=psahttp://www.jstor.org/stable/192623?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/192623?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=psahttp://www.jstor.org/action/showPublisher?publisherCode=ucpress
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    2/12

    John

    Stuart

    Mill,

    John

    Herschel,

    and the

    'Probability

    of

    Causes'

    John

    V.

    Strong

    Boston

    College

    1. Introduction

    Any discussion of

    probability

    theory

    in

    early

    nineteenth-century

    Bri-

    tain

    requires a

    specially

    vigorous

    exercise

    of the

    historical

    imagination.

    A

    century of

    subsequent

    thought

    about

    the

    foundations

    of

    the

    subject

    has

    made

    both

    historians and

    philosophers

    of

    science

    only

    too

    aware

    of

    howi

    elusive

    and

    polyvalent

    a

    term

    'probability'

    is;

    indeed,

    since

    Carnap,

    it

    has

    scarcely

    been

    possible

    to use the

    word

    without a

    qualifying

    adjective

    or

    subscript.

    Before

    the nineteenth

    century,

    on

    the other

    hand,

    during

    the

    'emergence'

    of formalized

    probabilistic

    thinking,

    writers from Pascal

    and

    Fermat

    onwards were

    keenly

    conscious that the

    concept

    they

    used

    was

    often ill-defined

    and

    equivocally

    employed.

    There

    is

    a

    sensation,

    therefore,

    of

    sailing

    unexpectedly

    into a

    patch

    of

    flat

    calm when one turns

    to such

    early

    Victorian

    writers

    (I

    use

    the

    term in

    its

    common,

    extended sense

    to refer

    to the

    period

    from

    around

    1830

    till the

    middle of the

    century)

    as

    Lubbock,

    Drinkwater-Bethune,

    Galloway,

    and,

    above

    all,

    De

    Morgan.2

    For

    them,

    probability

    is

    Laplace's

    Essai

    phil-

    osophigue [16]

    and

    Th4orie

    analytique

    [17]

    with

    such

    semi-popularized

    pre-

    sentations

    of

    Laplace's

    ideas

    as

    Lacroix's

    Traite e"lementaire

    [15]

    and

    the

    writings of

    Quetelet

    ([23

    ]; [24

    ],

    English

    translation

    [25 ]).

    Even a

    mathematician

    of

    the

    stature of

    De

    Morgan

    devotes

    virtually all

    his ef-

    forts in

    this

    field to

    making

    Laplace's

    results

    accessible

    to

    English-

    speaking

    readers.

    Given

    the

    state of

    British

    mathematics

    during

    these

    years,

    this is

    not

    too

    surprising.

    Only a

    short

    while

    had

    passed

    since

    the

    subject had

    been

    wakened

    from

    its

    slumbers

    by the

    labors

    of

    the

    Young

    Turks

    of the

    Cambridge

    'Analytical

    Society':

    John

    Herschel,

    George

    Peacock,

    Charles

    Babbage and

    their

    contemporaries.

    Much

    of

    the

    best

    British

    mathematics was

    still an

    import from

    the

    Continent

    (and

    indeed

    the

    early

    efforts

    of the

    Cambridge

    group

    had

    gone

    largely

    into

    translating

    standard

    French

    treatises).

    In

    PSA 1978, Volume 1, pp. 31-41

    Copyright

    (

    1978

    by

    the

    Philosophy

    of

    Science

    Association.

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    3/12

    32

    probability

    theory (as

    in celestial

    mechanics),

    Laplace

    towered

    over

    his

    peers,

    and

    hence the obvious

    task for

    the British

    innovators

    was the

    ex-

    position

    and

    dissemination

    of

    his ideas

    and, as their confidence

    grew,

    the further

    development

    and

    application

    of

    them.

    Both precedent (in

    the

    form of Laplace's ownworks, as well as those of earlier writers like Con-

    dorcet3),

    and

    acquiescence

    in Laplace's

    serene

    assurance

    that

    the theory

    as a whole

    was no

    more

    than "good

    sense

    reduced

    to calculation,"

    encour-

    aged such

    steps;

    after

    all, who would

    be so perverse

    as

    to insist

    that

    his

    field

    of interest

    required no

    common

    ense?

    At this

    point

    events

    took a

    turn of

    a kind familiar

    in the history

    of

    ideas.

    Just as

    in the

    seventeenth

    century

    a much-simplified

    version

    of

    the philosophy

    of

    Bacon's

    Novumorganum

    was adopted

    as 'Baconianism',

    and

    just

    as

    doctrines

    about

    nature and

    scientific

    method

    far

    less

    nu-

    anced than Newton's

    own views

    became

    the 'Newtonianism'

    of

    the Enlighten-

    ment;

    so

    Laplace's

    richly

    ambivalent

    view of

    the

    subject matter

    of the 'doc

    trine of chances' became, in the nineteenth century, a muchmore univocal

    'Laplacean'

    interpretation

    of probability.

    "Chance,"

    says De Morgan

    in an

    early

    essay, "is

    merely

    an expression

    of our ignorance

    of the chain

    of

    events

    which have

    led to

    any

    particular

    occurrence."([7],

    p.

    103).

    In

    modern parlance,

    probabilities

    are

    rational

    degrees

    of belief,

    and

    any

    question

    of

    probability

    is

    simply

    "an inquiry

    into

    the number

    of

    ways in

    which,

    under given

    conditions,

    an event may

    happen

    or fail";

    when

    these

    are enumerated

    "the proportion

    of these two

    is that of

    the

    probabilities

    for and against."(ibid.).

    The ground

    is thus

    laid for

    a

    critic

    dissatis-

    fied with

    the received

    interpretation

    of Laplace's

    ideas,

    to

    question

    whe-

    ther Laplace

    himself

    was

    a

    'Laplacean'

    in this

    sense;

    and as

    we shall

    see,

    this is in fact what happens.

    When

    these

    criticisms

    did develop they

    were (as might

    be anticipated)

    directed against

    both the

    presuppositions

    of

    the

    Laplacean

    doctrine and

    its

    applications.

    That some objections

    were based

    simply

    upon

    misconceptions

    or bias is clear

    from

    the fact

    that writers

    like

    De

    Morgan

    took

    pains

    on

    repeated

    occasions

    (see

    [7], [8],

    [9])

    to

    insist

    that the

    theory

    was nei-

    ther

    irreligious

    nor an inducement

    to

    gambling

    Other criticisms

    were

    aimed chiefly

    at

    the

    applications

    of

    the

    theory:

    for

    instance,

    whether

    in

    particular

    cases

    like

    the calculations

    of Condorcet

    on

    judicial

    verdicts,

    the

    mathematics

    had

    not

    run

    away

    from its

    users,

    giving

    impressively

    pre-

    cise but totally unrealistic conclusions.4 One line of criticism, though--

    that

    which

    challenged

    the

    legitimacy

    of the so-called

    'inverse'

    use of

    the

    probability

    calculus

    to estimate

    which

    of

    several

    causes,

    each

    capable

    of

    producing

    an observed

    effect,

    had in fact done

    so--bore

    much

    more

    directly

    on

    the fundamental assumptions

    of the

    Laplacean

    interpretation.

    As

    I

    have

    argued

    elsewhere

    [29]

    ,

    the attacks of

    George

    Boole (most

    notably

    in

    his

    The Laws of

    Thought

    [4])

    and

    of

    John

    Venn

    (in

    The

    Logic

    of

    Chance

    [31])

    drastically

    undermined

    the

    whole

    project

    of

    merging

    degree-of-belief

    ac-

    counts

    of

    probability

    with the

    method

    of

    hypothesis,

    to

    produce

    a

    probabil-

    istic

    logic

    of

    scientific

    inference.

    Indeed,

    after

    the

    appearance

    in

    1874

    of

    that

    curiously

    anachronistic

    work,

    Jevons'

    The

    Principles

    of

    Science

    [13] ,

    such

    efforts were almost

    entirely

    abandoned

    for

    nearly

    half

    a

    cen-

    tury, until revived (in a variety of very different guises) by Keynes [14],

    Reichenbach [27],

    Ramsey [26],

    and

    Carnap

    [63.5

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    4/12

    33

    Though a

    really

    adequate

    account

    of

    the

    developments

    noted in the

    last

    paragraph

    has

    yet to

    be

    written,

    the

    major events,

    and

    the

    identities

    of

    the

    principal

    contributors,

    seem

    well

    enough

    established.6

    What

    I

    should

    like

    to

    argue

    here is

    that,

    even

    as

    a

    crude

    sketch,

    the

    picture

    I

    have

    drawn is flawed, omitting as it does (and as does every discussion ofthe

    topic

    of which

    I am

    aware),

    one

    of the most

    interesting

    treatments

    of

    probability, both 'pure'

    and

    applied,

    to

    be

    found

    in

    the

    nineteenth

    cen-

    tury.

    I

    refer

    to the

    following

    root-and-branch

    attack on

    Laplace, which

    begins with this

    gloss on the

    initial

    definition of

    probability

    found

    in

    the

    Essai

    philosophique:

    If his

    [i.e.,

    Laplace's]

    unrivaled

    command

    ver the

    means which

    mathematics

    supply for

    calculating

    the results

    of

    given data,

    necessarily

    implied an

    equally

    sure judgment

    of

    what the data

    ought

    to

    be,

    I

    should

    hardly

    dare

    give utterance to

    my

    conviction,

    that

    in

    this

    opinion he

    is

    entirely

    wrong;

    that his

    foundation

    is

    altogether insufficient for the superstructure erected upon it;

    and that

    there

    is

    implied, in all

    rational

    calculation of

    the

    prob-

    abilities of

    events, an

    essential

    condition,

    which

    is

    either over-

    looked

    in Laplace's

    statement, or

    so

    vaguely

    indicated as

    neither

    to

    be suggested

    to the

    reader, nor

    kept in

    view by

    the writer

    him-

    self. ([21],

    pp.

    1140-1141).

    The

    author of this

    passage is John

    Stuart

    Mill; the

    book

    is the System

    of

    Logic. But

    these

    views

    were

    likely

    not familiar

    to

    many of Mill's con-

    temporaries,

    since they appeared

    only in the

    first

    (1843)

    edition

    of the

    Logic,

    being

    replaced in

    all

    editions from

    1846

    onwards with

    what

    was (or

    could at least easily be read as) a muchmore sympathetic appreciation of

    the

    standard

    'Laplacean'

    account

    of

    probability. What

    I would

    like to do

    in

    the

    remainder

    of this

    paper, after

    describing

    briefly

    Mill's

    position

    in the

    first edition of

    the Logic

    (hereafter SLI),

    is

    suggest some

    answers

    to the

    following

    questions: (1)

    Why s

    Mill

    talking

    about

    probability at

    all in

    the

    Logic? (2) What is

    the

    source of

    his

    initial, very

    negative

    views

    on the

    received

    interpretation?

    (3) Whydid

    these

    views alter

    so

    markedly in the

    interval between

    1843

    and 1846?

    and (4) What

    was

    his final

    position on the

    matter?

    Finally, I

    shall

    indicate why

    the

    whole subject

    seems to

    me not

    only

    interesting

    in

    itself,

    but

    also

    significant

    for under-

    standing how British

    philosophy

    in

    the

    nineteenth

    century tried to

    come to

    grips with the

    alluring but

    recalcitrant

    'doctrine

    of chances

    2.

    Probability

    and the

    Structure of

    the System of

    Logic

    For those

    accustomed

    to think

    of

    the

    relationship

    between

    probability

    and

    scientific

    inference in

    twentieth-century

    terms, the fact

    that Mill

    devotes

    three

    chapters of the

    System

    of

    Logic (in

    all

    editions) to the sub-

    ject

    of

    probability

    may

    come as a

    surprise.

    Unlike,

    say,

    Jevons,

    Mill al-

    lows

    only

    a

    minor

    and

    subsidiary

    role

    to

    hypotheses in science. For

    Mill,

    scientific

    explanation

    is,

    broadly

    speaking,

    subsumption

    under causal

    laws,

    with the

    latter

    interpreted as

    invariable sequences

    of

    phenomena

    ([21],

    pp.

    323-369);

    the

    logic of

    hypotheses is a

    set

    of

    procedures for

    eliminating

    all but one candidate among a set of laws, all of which purport to cover

    the

    same

    phenomena.

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    5/12

    34

    Probability

    does find

    a natural

    place in Mill's

    discussion,

    however.

    Central

    to Mill's treatment

    of causality

    (and

    hence to the

    argument

    of

    the Logic

    as a

    whole) is the

    doctrine

    of the plurality

    of causes

    ([21],

    pp. 434-453).

    Events are commonly

    he effect

    of an

    intersection

    of

    causal chains,

    of a

    'collocation'

    of causes.

    In order to

    find if

    there

    is in any given case a deeper order underlying such collocations, we re-

    quire

    means for

    deciding whether

    observed

    regularities

    amongphenomena

    are in fact mere

    coincidences,

    or

    whether

    they betray

    the

    presence of

    hitherto

    unsuspected

    causes at work.

    Now this

    is an issue

    which

    had already been

    raised by several

    genera-

    tions

    of writers.

    Originally discussed

    in connection

    with a kind

    of ar-

    gument

    to

    design in the

    early

    eighteenth

    century

    by John Arbuthnot

    (see

    [11],

    pp.

    166ff.), it emerged in

    a

    full-blown,

    secularized

    form in a

    1735

    essay

    by Daniel

    Bernoulli

    on the origin

    of the solar

    system

    ([3]; cf.

    r30],

    pp. 222-224).

    The

    question,

    used

    as an

    illustration

    by

    virtually

    every writer on probability for the next hundred and fifty years, was this:

    given

    the fact

    that

    all the known

    planets

    revolve in

    the same

    sense about

    the

    sun,

    and

    that

    their orbits

    all lie more

    or

    less in a plane,

    it seems

    sensible to

    ask whether

    the

    observed

    order could

    have

    arisen simply

    by

    co-

    incidence,

    rather

    than from

    some single

    cause.

    In other

    words, prior

    to

    theorizing

    about the

    mechanismwhich

    might

    account for

    the

    regularity,

    one

    ought

    first to ask

    whether

    there

    need

    be any

    mechanism.

    This question

    about

    the

    'probability

    of causes'--in

    Mill's

    language,

    the problem

    of

    "how to distinguish

    coincidences

    which are

    casual

    from

    those

    which

    are the result

    of law"

    ([211,

    p. 1148)--is

    clearly crucial

    to

    any account of scientific explanation of the sort Mill proposes. And in

    fact Mill

    accepts

    the

    validity

    of

    the calculation

    done

    by

    Daniel

    Bernoulli

    (and, following

    him in

    essentials,

    by Laplace).

    What Mill

    objects

    to strong

    ly

    in

    ELa

    is the implication

    that

    somehow such

    an assessment of

    probabili-

    ties

    can

    go beyond

    what is

    given

    in

    experience;

    "that

    there can be

    a

    mea-

    surement

    of probability

    where

    there

    is no

    experience."(

    [21],

    p.

    1143).

    For Mill,

    Laplacean probability

    calculations

    are valid

    so

    far,

    but

    only

    so

    far,

    as their foundations

    are

    empirical:

    "Conclusions

    respecting

    the

    prob-

    ability

    of

    a fact

    rest not

    upon

    a

    different,

    but

    upon

    the

    very

    same

    basis,

    as conclusions respecting

    its

    certainty;

    namely,

    not our

    ignorance,

    but

    our

    knowledge:

    knowledge

    obtained

    by experience,

    of

    the

    proportion

    between

    the cases

    in

    which

    the

    fact

    occurs,

    and

    those

    in which

    it does

    not occur.

    Every calculation of chances is grounded upon an induction..." ([21], p.

    1142; emphasis

    added).

    It

    seems

    quite

    clear from

    this

    passage

    (and

    the

    impression

    is

    further strengthened

    when it

    is

    taken

    in

    context)

    that

    Mill

    is

    arguing

    for a

    frequency

    interpretation

    of

    probability,

    almost

    a

    quarter

    century

    before

    Venn.

    I think

    the likeliest

    reason

    for

    this,

    and for

    Mill's

    stand

    against

    the

    near-unanimous

    acceptance

    of

    Laplaceanism

    by

    his

    contempo-

    raries,

    is

    implicit

    in what

    has

    already

    been

    said

    about the

    depth

    of

    his

    empiricist

    commitments

    though

    'empiricism'

    was

    not for him an

    attractive

    label).

    The received

    version

    of

    probability

    smacked

    too much

    of

    apriorism,

    of

    a

    device

    whereby

    truths about

    the world could be

    spun

    out

    of the

    human

    mind without

    observation

    or

    experiment,

    and

    "by

    a

    system

    of

    operations

    upon

    numbers, our ignorance...coined into a science."(ibid.).8

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    6/12

    35

    Nonetheless,

    I would like

    to

    suggest

    one source

    of

    influence on

    Mill's

    ideas that

    may

    have

    been

    of

    significance.

    The writer to whomMill

    seems

    to have turned most regularly for

    methodological inspiration

    is the

    Scot-

    tish Common

    ense philosopher Dugald Stewart.9 Stewart

    has

    relatively

    little to say on probability, but he does devote one long note to the sub-

    ject

    in

    his Philosophy

    of the

    Active and Moral

    Powers of

    Man

    (1828).

    A

    large part of this note is

    given

    over to

    quotations from

    a

    paper

    of Pre-

    vost

    and L'Huillier

    (1796)--characterized

    by

    Stewart

    as

    "very

    able"-- in

    which the authors assert that

    our

    judgments about 'Laplacean'

    probabili-

    ties, and those about observed

    causal regularities among

    natural events,

    "dependent de deux ordres

    de

    facultes

    differents":

    in

    particular, that

    the latter

    rely wholly on

    "la

    liason des

    idees'"lfound

    in

    our experience

    ([28], VII,

    pp.

    116-118). Given

    the paucity of criticism of

    the founda-

    tions of

    probability prior to

    Mill's time, and the accessibility of this

    passage

    to Mill while he was

    writing

    the

    Logic,

    the

    arguments

    of Prevost

    and

    L'Huillier could have

    supported,

    even if

    they did not

    inspire,

    Mill's

    misgivings about the Laplacean view.10

    3. Herschel, Mill, and the Second Edition

    of

    the

    System of

    Logic (1846)

    If the

    above account

    is even

    roughly correct, especially

    in its

    empha-

    sis on the ease

    with

    which

    a

    non-Laplacean, frequency

    interpretation

    of

    probability finds a place in Mill's

    philosophy of science, it becomes all

    the more

    surprising

    to find him

    apparently turning away sharply from

    this

    interpretation only three years

    later, in the 1846 edition of the Logic

    (?P2).

    "The

    theory

    of

    chances,"

    he

    writes

    there,

    "as conceived

    by

    Laplace

    and by

    mathematicians generally,

    has not the fundamental

    fallacy which

    I

    had ascribed to it."([211, p. 535). And on another point on which he had

    taken issue with

    Laplace

    in

    SLI

    (the

    rules for estimating the credibility

    of

    witnesses

    testifying

    to the

    occurrence

    of

    an

    improbable

    coincidence),

    Mill now

    writes, "This argument of Laplace's, though I

    formerly thought

    it

    fallacious,

    is

    irrefragable

    in the

    case which he supposes..."

    ([21],

    p. 636).

    Fortunately

    for

    the

    historian,

    it

    is

    possible to trace

    in

    great detail

    the

    stages

    in

    Mill's

    change

    of heart.ll

    There is abundant

    evidence that,

    for

    a writer

    preparing

    a

    treatise

    of

    unprecedented scope

    on the

    'methods

    of

    scientific

    investigation' (as

    the

    subtitle

    of the

    Logic

    ran),

    Mill

    knew painfully little about the natural sciences of his day.12 Even after

    the

    publication of

    ?L1,

    he

    was

    conscious

    that

    the

    book might

    well contain

    errors

    or

    omissions that he

    had

    been

    unable

    to detect; and

    accordingly,

    when Sir

    John

    Herschel wrote

    a

    cordial letter of thanks for

    a presentation

    copy of

    the Logic which Mill had

    sent him, the latter made bold to ask if

    Herschel

    would suggest any corrections that seemed

    necessary.13

    The

    choice

    of

    critic was

    a

    shrewd

    one.

    If

    Mill knew surprisingly

    lit-

    tle

    science, Herschel had

    mastered an

    astonishingly large

    amount.

    (His

    on-

    ly

    rival

    would

    have been William

    Whewell,

    with whose

    Philosophy

    of

    the In-

    ductive

    Sciences

    Mill

    had

    dealt

    quite sharply

    in the

    Logic.)

    Even

    more im-

    pressive than Herschel's accomplishments

    was his reputation;

    as

    one

    recent

    commentator has put it, for the early Victorians, to be scientific was to

    be as

    much

    like

    Herschel

    as

    possible ([5], p. 219). Though five

    years

    re-

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    7/12

    36

    turned

    from

    his epochal

    survey of the southern

    skies

    at the Cape of Good

    Hope,

    Herschel

    was still deep

    in the monumental

    task of preparing his

    ob-

    servations

    for publication; but

    he readily

    turned his attention

    to the

    Logic when

    Mill began serious

    work on a new

    edition in 1845. In

    December

    of that year, Herschel sent a long letter to Mill, detailing his objec-

    tions to Mill's strictures

    against

    Laplace's statement

    of the theory

    of

    probabilities, as presented

    in

    SL1. "With these objections

    I can

    no ways

    agree,"

    he wrote,

    "and I

    will

    not conceal

    from you that

    I read themwith

    great concern and an

    earnest wish

    that you would give

    them a full

    recon-

    sideration."14 Granted,

    says

    Herschel, that Laplace

    shares the

    tendency

    of many

    mathematicians

    sometimes

    "to prefer

    mathematical calculation

    by

    rule and formula to

    the plain

    exercise of common ense."

    But this

    is no

    more than

    a lapse on Laplace's

    part.

    Mill's

    criticism misses the

    point

    that

    Laplace

    was aware that only

    through experience could

    the formal

    cal-

    culus of probabilities--which

    is, as Herschel notes,

    "simply the

    theory of

    combinations"--be linked with

    the real world.

    Probability is, to

    be sure,

    epistemic, "a matter of opinion & judgment" that certain events are equal-

    ly possible

    "as regards our limited

    knowledge

    or conceptions."

    But the

    basis of all such assessments

    is experience, whether

    in the form

    of the di-

    rect

    evaluation

    of

    relative

    frequencies,

    or of

    more generalized

    and tacit

    use

    of observation and experiment.

    "How do we know,"

    asks Herschel,

    "that

    in a forest

    of

    trees

    half

    of

    which are oaks

    it

    is equally possible

    to stum-

    ble on an

    oak and another

    tree?

    Has anybody

    ever tried it?" And

    yet such

    a judgment is

    in no sense

    a

    priori:

    "A

    total absence

    of all knowledge

    of

    the connexion

    of

    events--or

    so to speak,

    of

    the mechanism

    of the

    events

    is incompatible

    with that state

    of

    mind which leads us

    to assert an equal

    probability...

    .We must

    see enough

    of

    the case

    to

    get

    or make

    for

    ourselves

    an impression (perhaps an erroneous one) that there does exist a similarity

    of

    circumstances.

    In short

    an

    opinion

    or a

    judgment

    to

    be worth the name

    must be grounded

    on

    something

    not

    merely negative."

    The 1846

    revision

    of the

    Logic incorporates

    all these

    criticisms.

    After

    summarizing

    his

    original

    view

    (with

    an

    amplified quotation

    from

    Laplace's

    Essai

    to

    open

    the chapter),

    Mill

    devotes several

    pages

    to a detailed

    recan-

    tation

    of his

    position

    in

    SLI.

    Probability

    is

    indeed

    "probability

    to

    us,"

    that

    is,

    the

    warranted

    degree

    of

    expectation

    we have that an event

    will

    occur

    ([21],

    p. 535).

    This

    degree

    of

    expectation

    can

    be

    calculated

    by

    the

    "common

    heory"

    of

    probability

    even

    in

    those

    cases in which

    we are

    unable

    to appeal to any "special experience"; under such conditions,

    the

    "logical

    ground

    of the

    process

    [of calculation]

    is our

    knowledge,

    such

    knowledge

    as

    we

    then

    have,

    of the laws

    governing

    the

    frequency

    of occurrence

    of the

    dif-

    ferent

    cases;

    but

    in

    this

    case

    our

    knowledge

    is limited to that

    which, being

    universal

    and

    axiomatic,

    does

    not

    require

    reference

    to

    specific

    experience,

    or to

    any

    considerations arising

    out of the

    special

    nature

    of the

    problem

    under

    discussion."([21],

    p. 537).

    4.

    Conclusion

    It

    is

    clear

    that

    one

    wants

    to ask at least

    two

    questions

    about

    Mill's

    revised

    estimate

    of the

    Laplacean (or perhaps

    Laplacean-Herschelian)

    ac-

    count of probability: is it fairer to Laplace's own views, and is it a

    valid

    interpretation

    of the doctrine

    of

    chances

    in its own

    right?

    The

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    8/12

  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    9/12

    38

    6In America

    the project

    was going

    forward,

    in a

    quite

    different

    form,

    in

    the work

    of C.S.

    Peirce.

    7Mill's

    criticisms

    in SL1

    may

    thus antedate

    those of Robert

    Leslie

    El-

    lis, whose brief, -difficult papers in the Transactions of the Cambridge

    Philosophical

    Society

    for 1843 and

    1856

    also

    argued for

    the

    frequency

    interpretation.

    8Hacking ([10],[11])

    has

    shown

    that Laplace's

    views

    uneasily

    embrace

    both degree-of-belief

    and

    relative-frequency

    accounts

    of probability,

    and that this

    duality can

    be traced

    back

    to the writings

    of James

    Bernoul-

    li and Leibniz

    on the

    subject.

    9See

    Olson

    ([22],

    esp. pp.

    94-124)

    for an

    extensive

    discussion

    of

    Stew-

    art's views

    and

    of their

    influence

    on Victorian

    philosophy

    of science.

    1OThe

    importance

    of this

    Continental

    tradition

    for the

    overall

    develop-

    ment of

    methodology

    in Britain

    has

    been stressed

    in

    a

    recent paper

    by

    Lau-

    dan [18 ]

    llThis

    process

    has been

    noted

    by

    J.M. Rob-,on

    in his "Textual

    Introduction

    to

    [21];

    see

    esp. pp.

    lxxx-lxxxi.

    12Mill

    was engagingly

    aware

    of his

    limitations

    in

    this

    regard,

    as

    his

    re-

    marks

    on the

    matter

    in

    his Autobiography

    ([19],

    pp.

    133-134)

    testify.

    13Although I have concentrated here on what seem to me to be the chief

    points

    of Herschel's

    criticisms

    regarding

    Mill's views

    on

    probability,

    Herschel

    also provided

    a

    number

    of

    corrections

    on

    other

    topics;

    see

    Mill's

    letter

    to Herschel

    of

    28

    February

    1846

    ([20],

    p.

    695).

    14Herschel

    to

    Mill,

    22 December

    1845;

    manuscript

    (apparently

    a fair copy

    byr

    n

    amanuensis)

    at

    the

    Royal

    Society,

    London.

    I am

    grateful

    for

    permis-

    sion

    to

    quote

    from

    this

    letter,

    to which the

    Society

    holds

    copyright.

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    10/12

    39

    References

    [1] Baker, K.M. Condorcet: From Natural

    Philosophy

    to

    Social Mathe-

    matics. Chicago and London: University

    of

    Chicago Press, 1975.

    [2] Blake, R.M., Ducasse, Curt J.

    and Madden, E.H.

    Theories of

    Scientific Method: The Renaissance Through the Nineteenth

    Century.

    Seattle: University

    of

    Washington Press, 1960.

    [3] Bernoulli, D. "Recherches

    physiques

    et

    astronomiques...

    ."

    In

    Pieces

    qui

    ont remporte le prix

    double

    de l'Adad6mie

    royale

    des

    sciences

    en

    MDCCXXXIV. Paris: Imprimerie royale, 1735.

    Pages

    93-122.

    [4] Boole,

    G.

    An Investigation of the Laws

    of

    Thought.

    London:

    Macmillan, 1854.

    [5] Cannon,

    W.F. "John

    Herschel

    and

    the Idea

    of

    Science." Journal

    of

    the

    History

    of

    Ideas

    22(1961):

    215-239.

    [6] Carnap,

    R.

    Logical Foundations

    of

    Probability. 2nd ed.

    Chicago:

    University

    of

    Chicago Press, 1962.

    [7]

    De

    Morgan, A. "Quetelet on Probabilities." Quarterly Journal of

    Education IV(1832): 101-110.

    [8] ------------. "Theory

    of

    Probabilities." In

    Encyclopaedia

    Metro-

    politana. London: B. Fellowes, 1817-1845.

    [9]

    ------------.

    An Essay

    on

    Probabilities. London: Longman and

    Taylor, 1838.

    [10] Hacking, I. "Equipossibility Theories of Probability." British

    Journal

    for

    the Philosophy of

    Science 22(1971): 339-355.

    [11]

    ?-----------

    The Emergence of

    Probability. Cambridge: Cambridge

    University Press, 1975.

    [12] Herschel, J.F.W. Essays

    from the Edinburgh and Quarterly Reviews.

    London: Longman, Brown, Green, Longmans, and Roberts, 1857.

    [13]

    Jevons,

    W.S. The Principles of Science. 2 vols. London: Mac-

    millan, 1874.

    [14] Keynes, J.M. A Treatise on

    Probability. London: Macmillan, 1921.

    [15] Lacroix, S.F.

    Traite

    glementaire

    du calcul des probabilites.

    2Pme

    6d.

    Paris: Bachelier, 1822.

    [16] Laplace, P.S. Essai

    Philosophigue

    sur les

    Probabilites.

    (In-

    cluded as the "Introduction" to [17]. Pages V-CLXIX. English

    translation:

    A

    Philosophical

    Essay on Probabilities. (trans.)

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    11/12

    40

    F.W. Truscott

    and

    F.L. Emory. New York: J. Wiley

    & Sons,

    1902.

    Reprint

    by Dover,

    1951.)

    [17]

    --------------. Th6orie

    Analytique

    des Probabilites.

    3rd

    edi-

    tion. (1820). In Oeuvres de Laplace, Volume VII. Paris:

    Gauthier-Villars,

    1878-1912.

    [18]

    Laudan,

    L.

    "The

    Sources of

    Modern

    Methodology." In

    Historical

    and Philosophical

    Dimensions

    of Logic,

    Methodology

    and Philosophy

    of Science.

    (University

    of

    Western

    Ontario Series

    in Philosophy

    of

    Science, Volume

    12.)

    Edited by

    R.E.

    Butts and J.

    Hintikka.

    Dordrecht:

    D. Reidel,

    1977.

    Pages

    3-19.

    [19]

    Mill,

    J.S.

    Autobiography.

    Indianapolis:

    Bobbs-Merrill,

    1957.

    (Originally

    published 1873.)

    [20]

    ---------.

    The Earlier Letters of John Stuart Mill 1812-1848.

    2 volumes.

    Edited

    by F.E. Mineka.

    Toronto:

    University

    of

    Toronto Press.

    London:

    Routledge

    and Kegan Paul,

    1963.

    [21] ---------.

    A System

    of Logic

    Ratiocinative

    and

    Inductive.

    Volumes

    7 and 8 of

    the

    Collected

    Works of

    John Stuart

    Mill.

    Edited by J.M. Robson.

    Toronto:

    University

    of Toronto

    Press,

    1974.

    [22]

    Olson,

    R. Scottish

    Philosophy

    and British

    Physics

    1750-1880.

    Princeton:

    Princeton

    University

    Press,

    1975.

    [23]

    Quetelet, A.

    Instructions

    populaires

    sur le

    calcul

    des probabili-

    tes.

    Bruxelles:

    Tarlier et

    Hayez,

    1828.

    [24]

    -----------.

    Lettres...

    sur la

    th6orie

    des

    probabilit9s.

    Bruxelles:

    Hayez,

    1846.

    [25]

    -----------.

    Letters...

    .on the

    Theory

    of

    Probabilities.

    (trans.)

    O.G.

    Downes. London: Layton,

    1849.

    (A

    translation

    of

    [24].)

    [26]

    Ramsey,

    F.P. "Truth

    and

    Probability."

    In The

    Foundations

    of

    Mathematics and Other Logical Essays. (ed.) R.'B.

    Braithwaite.

    Paterson,

    NJ:

    Littlefield,

    Adams,

    &

    Co.,

    1960.

    Pages

    156-211.

    [27]

    Reichenbach,

    H.

    Wahrscheinlichkeitslehre.

    Leiden:

    Sijthoff,

    1935.

    (Translated

    and revised

    as The Theory

    of

    Probability.

    Berkeley:

    University

    of

    California

    Press, 1949.)

    [28] Stewart,

    D.

    The

    Collected

    Works

    of

    Dugald

    Stewart.

    Edited

    by

    Sir

    William Hamilton. Farnborough:

    Gregg

    International,

    1971.

    (A

    reprint

    of

    the

    Edinburgh

    edition of

    1854-1860.)

    [29]

    Strong,

    J.V.

    "The

    Infinite Ballot

    Box

    of Nature:

    De

    Morgan,

    Boole, and Jevons on Probability and the Logic of Induction."

    In PSA

    1976,

    Volume

    1.

    Edited

    by

    F.

    Suppe

    and P.D.

    Asquith.

    This content downloaded on Mon, 11 Feb 2013 19:34:32 PMAll use subject toJSTOR Terms and Conditions

    http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp
  • 8/11/2019 Strong. John Stuart Mill, John Herschel, And the Probability of Causes (Article)

    12/12

    41

    East Lansing, Michigan: Philosophy of Science Association,

    1976.

    Pages 197-211.

    [30] Todhunter, I. A History of the Mathematical Theory of Probabil-

    ity. London and Cambridge: Cambridge University Press, 1865.

    [31] Venn, J. The Logic of Chance. London and Cambridge:

    Macmillan,

    1866.