Sie sind auf Seite 1von 8

Strong Inference

John R. Platt

Science, New Series, Vol. 146, No. 3642. (Oct. 16, 1964), pp. 347-353.

Stable URL:
http://links.jstor.org/sici?sici=0036-8075%2819641016%293%3A146%3A3642%3C347%3ASI%3E2.0.CO%3B2-K

Science is currently published by American Association for the Advancement of Science.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained
prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in
the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/journals/aaas.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For
more information regarding JSTOR, please contact support@jstor.org.

http://www.jstor.org
Thu Apr 5 20:06:34 2007
16 October 1964, Volume 146, Number 3642 SCIENCE

"nature" o r the experimental outcome


chooses-to go to the right branch o r
the left; at the next fork, to go left
or right; and so on. There are similar
Strong Inference branch points in a "conditional com-
puter program," where the next move
depends on the result of the last cal-
culation. And there is a "conditional
Certain systematic methods of scientific thinking inductive tree" o r "logical tree" of this
may produce much more rapid progress than others. kind written out in detail in many
first-year chemistry books, in the table
of steps for qualitative analysis of a n
J o h n R. Platt unknown sample, where the student
is led through a real problen~of con-
secutive inference: Add reagent A; if
1011get a red precipitate, it is sub-
Scientists these days tend to keep in scientific advance is an intellectual group alpha and you filter and add
up a polite fiction that all science is one. These rapidly moving fields are reagent B; if not, you add the other
equal. Except for the work of the mis- fields where a particular method of reagent, B'; and so on.
guided opponent whose arguments we doing scientific research is systemati- On any new problem, of course,
happen to be refuting at the time, we cally used and taught, an accumulative inductive inference is not as simple
speak as though every scientist's field method of inductive inference that is and certain as deduction, because it
and methods of study are as good as so effective that I think it should be involves reaching out into the un-
every other scientist's, and perhaps a given the name of "strong inference." known. Steps 1 and 2 require in-
little better. This keeps us all cordial I believe it is important to examine tellectual inventions, which must be
when it comes to recon~mending each this method, its use and history and cleverly chosen so that hypothesis, ex-
other for government grants. rationale, and to see whether other periment, outcome, and exclusion will
But I think anyone who looks at groups and individuals might learn to he related in a rigorous syllogism; and
the matter closely will agree that some adopt it profitably in their own scien- the question of how to generate such
fields of science are moving forward tific and intellectual work. inventions is one which has been ex-
very much faster than others, perhaps In its separate elements, strong in- tensively discussed elsewhere (2, 3 ) .
by an order of magnitude, if numbers ference is just the simple and old- What the f o r n ~ a l schema reminds us
could be put on such estimates. The fashioned method of inductive infer- to do is to try to make these inven-
discoveries leap from the headlines- ence that goes back to Francis Bacon. tions, to take the next step, to proceed
and they are real advances in complex The steps are familiar to every college to the next fork, without dawdling o r
and difficult subjects, like molecular student and are practiced, off and on, getting tied up in irrelevancies.
biology and high-energy physics. As by every scientist. The difference comes Jt is clear why this makes for rapid
Alvin Weinberg says ( I ) , "Hardly a in their systematic application. Strong and powerful progress. For exploring
month goes by without a stunning suc- inference consists of applying the fol- the unknown, there is no faster meth-
cess in molecular biology being re- lowing steps to every problen~ in sci- od; this is the minimum sequence of
ported in the Proceedings of the Na- ence, formally and explicitly and regu- steps. Any conclusion that is not a n
tional Academy of Sciences." larly: exclusion is insecure and must be re-
Why should there be such rapid ad- 1 ) Devising alternative hypotheses; checked. Any delay in recycling to the
vances in some fields and not in others? 2) Devising a crucial experiment (or next set of hypotheses is only a delay.
I think the usual explanations that we several of them), with alternative possi- Strong inference, and the logical tree
tend to think of-such as the tracta- ble outcomes, each of which will, as it generates, are to inductive reasoning
bility of the subject, or the quality or nearly as possible, exclude one or more what the syllogism is to deductive rea-
education of the men drawn into it, of the hypotheses; soning, in that it offers a regi~larmeth-
or the size of research contracts-are 3) Carrying out the experiment so od for reaching firm inductive con-
important but inadequate. I have be- as to get a clean result; clusions one after the other as rapidly
gun to believe that the primary factor 1 ') Recycling the procedure, making hs possible.
subhypotheses or sequential hypotheses "But what is so novel zlhout this?"
T h e author is professor of hiopliysics and to refine the possibilities that remain; someone will say. This is rhe method
physics a t the University of Chicago, Chicago,
111. This is the text of a n address given before and so on. of science and always has been; why
the Division of Physical Chemistry of the Amer- give it a special name? The reason is
ican Chemical Society in September 1963. under
It is like climbing a tree. At the
the title "The New Raconians." first fork, we choose-or, in this case, that inany of us have alnlost forgotten
16 OCTOBER 1964
it. Science is now a n everyday business. ments John Dalton would have liked, chemist o r scientist in any field ac-
E,quipn~ent, calculations, lectures be- n h e r e the combining entit~cs are not customed to less closely reasoned arti-
come ends in themselves. H o w many atoms but long n~acronlolecular cles and less sharply stated inferences
of us u r l t e clown our alternatives and strands. will find it a salutary experience to
crucial experiments every day, Eocus- O r take a different sort of question: dip into that journal almost at random.
Ing on the e,xclu\ion of a hypothesis? Is the "genetic map"-show~ng the sta-
W e niay write our scientific papers so tistical relationship of different genetic
that it looks as if we had steps 1, 2, characteristics in recombination exper- Resistance to
and 3 in mind all along. But in be- iments-a one-dimensional map like Analytical Methodology
tween, we do busywork. W e become the DNA ~ilolecule (that is, a linear
"method-oriented" rather than "prob- map), as T. H . Morgan proposed in This analytical approach to biology
lem-or~ented." W e say we prefer to 191 1, o r does it have two-d~mensional has sometimes become almost a cru-
"feel our way" toward generalizations. loops or branches? Sey~ilour Benzer sade, because it arouses so much re-
W e la11 to teach our students how to showed that his hundreds of fine micro- sistance in many scientists who have
sharpen u p their inductive inferences. genetlc experi~ilents on b a c t e r ~ awould grown up in a more relaxed and diffuse
And wc d o not realize the added fit only the n~athematical matrix for tradition. At the 1958 Conference on
power that the regular and explicit the one-dimensional case ( 7 ) . Biophysics, at Boulder, there was a
use of alternative hypotheses and sharp I3ut of course, selected crucial ex- dramatic confrontation between the
exclusions could give us at every step periments o t this kind can be found two points of view. Leo Szilard said:
of our research. in every field T h e real difference in "The problems of how enzymes are
T h e differcnce between the average molecular biology is that tornial in- induced, of how proteins are synthe-
scientist's informal n ~ e t h o d s and thecIuctive inference is so systematically sized. of how antibodies are formed,
methods of the strong-interence users practiced and taught. O n any given are closer to solution than is generally
i\ so~ilewhat l ~ k e the difference be- rnornlng at the 1 aboratory of Molecu- believed. If you d o stupid experiments,
tween a gasoline engine that fires oc- lar Biology in Cambridge, England, and finish one a year, it can take 50
casionally and one that fires in steady the blackboards of Francis Crick or years. But if you stop doing experi-
sequence. If our motorboat engines Sidney Brenner will commonly b e ments for a little while and ihinlc how
were as ejrdtic as our deliberate in- found covered with logical trees. On proteins can possibly be synthesi~ed,
lellectr~al efforts, rvlost of us would not the top line will be the hot new result there are only about 5 different ways,
get home for supper. just u p from the laboratory o r just in not SO! And it will take only a few
hy letter o r rumor. On the next line experiments to distinguish these."
will be two o r three alternative ex- One of the young men added: "It
Molecular Biology planations, or a little list of "What h e is essentially the old question: How
did wrong." Underneath will be a se- snzall and elegant a n experiment can
'The new molecular biology is a ries o t suggested experiments or con- you perform?"
fielci where I thlnk this systematic trols that can reduce the number o t These comments upset a number of
nietliod of inference has become wide- possibilities. And so on. T h e tree grows those present. AII electron rnicroscopi5t
4pread and effective. It is a complex during the day as one m a n or another said. "Gentlemen, thi5 is off the track.
Field; yet a succession of crucial ex- comes in and argues about why one This is philosophy of science."
periments over the past decade has a t the experiments wouldn't work, o r Szilard retorted, "1 was not quarrel-
given us a surprisingly detailed under- how it should be changed. ing with third-rate scientists: 1 was
standing of hereditary mechanisms and T h e strong-inference attitude is evi- quarreling with first-rate scientists."
the control of enzynle formation and dent ju\t in the style and language in A physical chemist hurriedly asked,
protein synthesis. which the papers are written. F o r ex- "Are we going to take the official
T h e logical structure shows in every ample, in analyzing theories of anti- photograph before lunch o r after
cxperitiient. In 1953 James Watson body formation, Joshua Lederberg lunch?"
and Francis Crick proposed that the ( 8 ) gives a list of nine propositions But this did not cleflect the dispute.
DNA molecule-the "hereditary sub- "subject to denial," discussing which A distinguished cell biologist rose and
stance" in a cell---is a long two- ones would be ''most vulnerable to said, "No two cells give the same
stranded helical molecule ( 4 ) . This sug- experimental test." properties. Biology is the science of
gested a number o t alternatives tor T h e papers of the French leaders heterogeneous systems." And h e added
crucial test. D o the two strands of F r a n ~ o i s Jacob and Jacques Monod privately, "You know there are rci-
the helix stay together when a cell are also celebrated t o r their high "logi- entirts; and there arc people in science
divides, or do they separate? Matthew cal density," with paragraph aftcr para- who are just working with these over-
Mcselson and Franklin Stahl used an graph of linked "incl~ictive 5yllogisn1s." simplified model systems-DNA chains
ingenious isotope-den5ity-labeling tech- Brit the style is widespread. Start with and in vitro systems---who are not
nique which showed that they sepa- the first paper in the Journal o f Mo- ctoing science a t all. W e need their
rate (5). Does the D N A helix always lec~llarBiology f o r 1964 (9). and you auxiliary work: they build apparatus,
have t ~ 7 0 strands, o r can it have immediately find: "Our conclusions . . . they make nzinor studies, but they are
three, i i ~atomic niodels suggest? Alex- might be invalid if . . . (i) . . . (ii) not scientists."
ander Rich 5howed it can have either, . . . o r (iii). . . W e shall describe ex- Ts which Cy Levinthal replied:
depending on the ionic concentration periments which eliminate these al- "Well, there are two kinds of biolo-
(6). These are the kinds of experi- ternatives." T h e average physicist o r gists, those who are looking to see
SCIENCE, VOL. 146
if there is one thing that can be under- new particles explicitly enough so that from the new accelerators and the
stood, and those who keep saying it if they are not found the theories will million-dollar costs of operation have
is very complicated and that nothing fall. As the biologist W. A. H. Rush- forced a similar analytical approach.
can be understood. . . . You must ton has said ( I ] ) , "A theory which It pays to have a top-notch group
study the sirrlplest system you think cannot be mortally endangered cannot debate every experiment ahead of
has the properties you are interested be alive." hlurray Gell-Mann and time; and the habit spreads throughout
in." Yuval Ne'eman recently used the parti- the field.
As they were leaving the meeting, cle grouping which they call "The
one man could be heard muttering, Eightfold Way" to predict a n~issing
-'\?'hat does Szilard expect me to do- particle, the Omega-Minus, which was Induction and Multiple Hypotheses
shoot myself?" then looked for and found (12). But
Any criticism or challenge to con- one alternative branch of the theory Historically, I think, there have been
sider changing our methods strikes of would predict a particle with one-third two main contr~butions to the de-
course at all our ego-defenses. Hut the usual electronic charge, and it was velopnlent of a satisfactory strong-
in this case the analytical metl~od of- not found in the experiments, so this inference method. The first is that of
fers the possibility of such great in- branch must be rejected. Francis Bacon (13). H e wanted a
creases in efroctiveness that it is un- The logical tree is so much a "surer method" of "finding out nature"
fortunate that it cannot be regarded part of high-energy physics that some than either the logic-chopping or all-
more often as a challenge to learning stages of it are con~monly built, in inclusive theories of the time or the
rather than as a challenge to combat. fact, into the electronic coincidence laudable but crude attempts to make
Many of the recent triumphs in mo- circuits that detect the particles and lnductlons "by silnple enumeration."
lecular biology have In fact been trigger the bubble-chamber photo- H e did not merely urge experiments,
achieved on just such "oversimplified graphs. Each kind o i particle should as some suppose; h e showed the fruit-
model systems," very much along the give a different kind of pattern in the f ulness of interconnecting theory and
analytical lines laid down in the 1958 electronic counters, and the circuits can experiment so that the one checked
d~scussion.They have not fallen to the be set to exclude or include whatever tbe other. Of the many inductive pro-
kind of men who justify themselves types of events are desired. If the dis- cedures he suggested, the most im-
by saying, "No two cells are alike," tinguishing criteria are sequential, they portant, 1 think, was the condi-
regardless of how trut that mav ulti- may even run through a complete logi- tional inductive tree, which pro-
mately be. The triumphs are in fact cal tree in a microsecond or so. This ceeded from alternative hypotheses
triumphs of a new way of thinking. electronic preli~ilinaryanalysis. like hu- (possible "causes," as he calls them),
man preliminary analysis of alterna- through crucial experiments ("ln-
tive outcomes, speeds up progress by stances of the Fingerpost"), to exclu-
High-Energy Physics sharpening the criteria. It eliminates slon of some alternatives and adoption
hundreds of thousands of the irrele- of what is left ("establishing axioms").
This analytical thinking is rare, but vant pictures that formerly had to be HIS Instances of the Fingerpost are
it is by no means restricted to the scanned, and when it is carried to its explicitly at the forks in the logical
new biology. High-energy physics is limit, a few output pulses, hours apart, tree, the term being borrowed "from
another field where the logic of ex- may be enough to signal the existence the fingerposts which are set up where
clusions is obvious, even in the news- of the antiproton or the fall of a loads part, to indicate the several di-
paper accounts. For example, in the theory. rections."
fanlous discovery of C . N . Yang and 1 think the einphasis on strong in- Many of h ~ crucial
s experiments pro-
T . D . Lee, the question that was ference in the two fields I have men- posed in Book IT of T h e h e w Orgnnon
asked was: D o the fundamental parti- tioned has been partly the result of ale st111 fascinating. For example, in
cles conserve mirror-symmetry or "par- personal leadership, such as that of order to decide whether the weight of
ity" in certain reactions, or do they the classical geneticists in ~ilolecular a body is due to its "inherent nature,"
not? The crucial experiments were biology. or of Szilard with his "Mid- as some had said, or is due to the
suggested; within a few months they west Chowder and Bacteria Society" attraction of the earth, which would
were done, and conservation of parity at Chicago in 1948-50, or of Max decrease with distance, h e proposes
was found to be excluded. Richard Delbriick with his summer courses in comparing the rate of a pendulum
Garwin, Leon Lederman, and Marcel phage genetics at Cold Spring Harbor. clock and a spring clock and then
Weinrich did one of the crucial ex- But it is also partly due to the nature lifting them froin the earth to the top
periments. It was thought of one of the fields then~selves.Biology, with of a tall steeple. H e concludes that if
evening at suppertime; by midnight its vast informational detail and com- the pendulum clock on the steeple
they had rearranged the apparatus for plexity, is a "high-information" field. "goes more slowly than it did on ac-
it. and by 4 a.m. they had picked up where years and decades can easily count of the diminished virtue of its
the predicted pulses showing the non- be wasted on the usual type of "low- weights . . . we may take the attrac-
con5ervation of parity (10). The phe- information" observations or experi- tion of the mass of the earth as the
nomena had just been waiting, so to ments if one does not think carefully cause of weight."
s ~ e a k ,for the explicit forniulation of in advance about what the most im- Here was a method that could sepa-
the alternative hypotheses. portant and conclusive experiments rate off the empty theories!
The theorists in this field take pride would be. And in high-energy physics, Bacon said the inductive method
in trying to predict new properties or both the "information flux" of particles could be learned by anybody, just like
16 OCTOBER 1964
learning to "draw a straighter line or original explanation for a phenome- t h o ~ ~ g h t "that Chamberlin described,
more perfect circle . . . with the help riot1 which seems satisfactory, that mo- the reason for the sharpness, the ex-
of a ruler or- a pair of compasses." ment affection for his intellectual child citement, the zeal, the teamwork-yes,
"My way of discovering sciences goes springs into existence, and as the ex- even international teamwork-in mo-
Ear to level men's wit and leaves but planation grows into a definite theory lecular biology and high-energy phys-
!ittle to inclividual excellence, because his parental affections cluster about his ics today. What else could be so ef-
it pertorms everything b y the surest offspring and it grows more and more fective?
rules and demonstrations." Even oc- dear t o him. . . . There springs u p When niultiple hypotheses become
casional mistakes would not be fatal. also unwittingly a pressing of the the- coupled to strong inference, the sci-
"Truth will sooner come out from ory to make it fit the facts and a entific search becomes an emotional
crror than from confusion." pressing of the facts to make them powerhouse as well as an intellectual
Tt is easy to see why young minds fit the theory. . . . one.
leaped to try it. "To avoid this grave danger, the Unfortunately, 1 think, there are
Nevertheless there is a difficulty with lnethod of multiple working hypotheses other areas of science today that are
this method. As Bacon emphasizes, it is urged. It differs from the simple sick by cotnparison, because they have
is necessary to make "exclusions." H e working hypothesis in that it distributes forgotten the necessity for alternative
says. "The induction which is to be the effort and divides the affections. hypotheses and disproof. Each m a n
available for the discovery and dem- . . . Each hypothesis suggests its own has only one branch-or none-on the
onstration of sciences and arts, must criteria, its own means of proof, its logical tree, and it twists at random
analyze nature by proper rejections own lnethod of developing the truth, without ever coming to the need for
and cxclusions; and then, after a suffi- and if a group of hypotheses encom- a crucial decision at any point. W e
cient number of negatives, come to a pass the subject on all sides, the total can see from the external syn~ptotns
conclusion on the affirlnative in- outcome of means and of methods is that there is something scientifically
stances." "[To n-ian] ~t is granted only full and rich." wrong. The Frozen Method. The Eter-
to proceed at first by negatives, and Chaniberlin thinks the tiiethod "leads nal Surveyor. T h e Never Finished. T h e
at last to end in afirniatives after to certain distinctive habits of ~ n i n d " Great Man With a Single I-lypothesis.
exclusion has been exhausted." and is of prime value in education. T h e Little Club of Dependents. T h e
Or, as the philosopher Karl Popper "When faithfully followed for a sufi- Vendetta. T h c All-Encompassing The-
says today, there is n o such thing as cient time, it develops a mode of ory Which Can Never Be Falsified.
proof in science--because some later thought of its own kind which niay Some cynics tell a story, which niay
altertiative explanation may be as good be designated the habit of conlplex be apocryphal, about the theoretical
or better-so that science advances thought. . . ." chemist who explained to his class,
only hy disproofs. There is no point This charming paper deserves to be "And thus we see that the C-CI
in making hypotheses that are not reprinted in some more accessible bond is longer in the first compound
falsifiable, became such hypotheses do journal today, where it could be re- than in the second because the percent
riot say anything: "it must be possible quired reading for every graduate stu- of ionic character is stiialler."
for an empirical scientific systeni to dent-and for every professor. A voice from the back of the rootii
he rcfuted by experience" (14). It seems to m e that Chatiiberlin has said, "But Professor X, according to
T h e difiiculty is that disproof is a hit o n the explanation-and the cure the Table, the C-Cl bond is shorter
hard doctrine. If you have a hypothesis -for many of our problen~s in the in the first con~pound."
and 1 h'ive another hypothesis, evi- sciences. The conflict and exclusion "Oh, is it?" said the professor.
dently one of them must be eliminated. of alternatives that is necessary to "Well, that's still easy to understand,
The scientist seems to have no choice sharp inductive inference has been all because the double-bond character is
hut to be either soft-headed or dis- too often a conflict between men, each higher in that compound."
putatious. Perhaps this is why so many with his single Ruling Theory. But T o the extent that this kind of story
tend to resist the strong analytical ap- whenever each Inan begins to have is accurate, a "theory" of this sort is
proach--and why some great scientists multiple working hypotheses, it be- not a theory at all, because it does
are so disputatious. conies purely a conflict between ideas. not exclude anything. It predicts every-
Fortunately, it seems to me, this It beco~iiesmuch easier then for each thing, and therefore does not predict
difficulty can be retnoved by the use of us to aitii every day at conclusive anything. It becotiies simply a verbal
of a second great intellectual invention. disproofs-at .strong inference-with- formula which the graduate student
the "tiiethod of niultiple hypotheses," out either reluctance or cotiibativeness. repeats and believes because the pro-
which is what was needed t o round I n fact, when there are multiple hy- fessor has said it so often. This is not
out the Baconian scheme. This is a potheses which are not anyone's "per- science. but faith; not theory, but
rnethod that was put forward by T. C . sonal property" and when there are theology. Whether it is hand-waving
Chamberlin (15), a geologist at Chi- crucial experiments t o test them, the or number-waving or equation-waving,
cago at the turn of the century, who daily life in the laboratory takes on a theory is not a theory unless it can
is best known for his contribution to a n interest and excitement it never be disproved. That is, unless it can
the Chamberlin-Moulton hypothesis of had, and the students can hardly wait be falsified by some possible experi-
the origin of the solar systeni. to get to work to see how the de- mental outcome.
Chaniberlin says our trouble is that tective story will come out. It seetiis In chemistry, the resonance the-
when we make a single hypothesis, to m e that this is the reason for the orists will of course suppose that I
wc become attached to it. cleveloptiient of those "distinctive hab- a m criticizing then?. while the molecu-
"The moment one has offered an its of mind" and the "complex lar-orbital theorists will suppose I am
SCIENCE, VOL. 146
criticizing tlzelii. But their actions- wine and beer, to the disease of silk- Roentgen, to see how fast we can
our actions, for I include myself among worms, to the problem of "spontane- pass from the general survey to ana-
them-speak for themselves. A failure ous generation," to the anthrax disease lytical inferences. We should try, like
to agree for 30 years is public ad- of sheep, to rabies. In each of these Pasteur, to see whether we can reach
vertisement of a failure to disprove. fields there were experts in Europe strong inferences that encyclopedism
My purpose here, however, is not who knew a hundred times as much could not discern.
to call names but rather to say that as Pasteur, yet each time he solved We speak piously of taking mea-
we are all sinners, and that in every problerns in a few months that they surements and nuking small studies
field and in every laboratory we need had not been able to solve. Obviously that will "add another brick to the
to try to formulate multiple alterna- it was not encyclopedic knowledge that te~nple of science." Most such bricks
tive hypotheses sharp enough to be produced his success, and obviously just lie around the brickyard (20).
capable of disproof. it was not simply luck, when it waq Tables of constants have their place
repeated over and over again; it can and value, but the study of one spec-
only have been the systematic power trum after another. if not frequently
Systematic Application of a special method of exploration. re-evaluated, may become a substitute
Are bacteria falling in'? Make the necks for thinking, a sad waste of intelli-
I think the work methods of a of the flasks S-shaped. Are bacteria gence in a research laboratory, and a
number of scientists have been testi- sucked in by the partial vacuum? Put mistraining whose crippling effects may
mony to the power of strong inference. in a cotton plug. Week after week his last a lifetime.
Is success not due in many cases to crucial experiments build up the logical To paraphrase an old saying, Be-
systematic use of Bacon's "surest rules tree of exclusions. The drama of ware of the man of one methcd or
and demonstrations" as much as to strong inference in molecular biology one instrument, either experimental or
rare and unattainable intellectual pow- today is only a repetition of Pasteur's theoretical. He tends to become ~neth-
er? Faraday's famous diary ( 1 6 ) , or story. od-oriented rather than problem-ori-
Fern~i's notebooks (3, 1 7 ) , show how The grand scientific syntheses, like ented. The method-oriented man is
these men believed in the effectiveness those of Newton and Maxwell, are shackled; the problem-oriented man is
of daily steps in applying formal in- rare and individual achievements that at least reaching freely toward what is
ductive methods to one problem after stand outside any rule or method. most important. Strong inference re-
another. Nevertheless it is interesting to note directs a man to problem-orienta-
Within 8 weeks after the discovery that several of the great synthesizers tion, but it requires him to be willing
of x-rays, Roentgen had identified 17 have also shown the strong-inference repeatedly to put aside his last methods
of their major properties. Every stu- habit of thought in their other work, and teach himself new ones.
dent should read his first paper (18). as Newton did in the inductive proofs On the other hand, I think that
Each demonstration in it is a little of his Opticks and Maxwell did in anyone who asks the question about
jewel of inductive inference. How else his experimental proof that three and scientific effectiveness will also con-
could the proofs have gone so fast, only three colors are needed in color clude that much of the n~athematiciz-
except by a method of maximum ef- vision. ing in physics and chemistry today is
fectiveness? irrelevant if not misleading.
Organic chemistry has been the The great value of mathematical
spiritual home of strong inference A Yardstick of Effectiveness formulation is that when an experi-
from the beginning. D o the bonds al- ment agrees with a calculation to fivz
ternate in benzene or are they equiva- I think the evident effectiveness of decimal places, a great many alterna-
lent? If the first, there should be five the systematic use of strong inference tive hypotheses are pretty well ex;-
disubstituted derivatives; if the second, suddenly gives us a yardstick for think- cluded (though the Bohr theory and
three. And three it is (19). This is a ing about the effectiveness of scientific the Schrodinger theory both predict
strong-inference test-not a matter of n~ethods in general. Surveys, taxon- exactly the same Rydberg constant!).
measurement, of whether there are omv, design of equipment. systematic But when the fit is only to two deci-
grams or milligrams of the products, measurements and tables, theoretical mal places, or one, it may be a trap
but a matter of logical alternatives. con~putations-all have their proper for the unwary; it may be no better
How else could the tetrahedral carbon and honored place, provided they are than any rule-of-thumb extrapolation,
atom or the hexagonal symmetry of parts of a chain of precise induction and some other kind of qualitative ex-
benzene have been inferred 50 years of how nature works. Unfortunately, clusion might be more rigorous for
before the inferences could be con- all too often they become ends in testing the assunlptions and more im-
firn~edby x-ray and infrared measure- themselves, mere time-serving from the portant to scientific understanding than
ment? point of view of real scientific ad- the quantitative fit.
We realize that it was out of this vance, a hypertrophied methodology I know that this is like saying that
kind of atmosphere that Pasteur came that justifies itself as a lore of respect- the emperor has no clothes. Today we
to the field of biology. Can anyone ability. preach that science is not science un-
doubt that he brought with hi111 a We praise the "lifetime of study," less it is quantitative. We substitute
completely different method of reason- but in dozens of cases, in every field, correlations for causal studies, and
ing? Every 2 or 3 years he moved to what was needed was not a lifetime physical equations for organic reason-
one biological problem after another, but rather a few short months or ing. Measurements and equations are
from optical activity to the fermenta- weeks of analytical inductive inference. supposed to sharpen thinking, but, in
tion of beet sugar, to the "diseases" of [n any new area we should try, like my observation, they more often tend
16 OCTOBER 1964
to nlakc the thinking noncausal and and learned. The nlolecular biologists and nuclear physics the response to
ttizry. l h c y tend to become the object today are living proof of it. T h e sec- T h e Question would bc to outline im-
ol s c ~ e n t ~ f i m
c anipulation instead of ond thing IS to be explicit and toririal niediately not one but several tests to
alix111ar) te\ts of crucial inferences. and regular about it, to devote a half disprove the hypothesis- -and it would
M a n j -perhaps niost--of the great hour or an hour to analytical t h ~ n k i n g turn out that the speaker already had
iss~tes of sclcnce are qualltat~ve, not every day, writing out the logical tree two or three graduate students work-
quantltritive, even In phy5ics and cheni- and the alternatives and ctucial experl- ing on them!
istty. Ikjuauons and measurements are nlents explicitly in a permanent note- I almost think that government agen-
useful \+hen drrd only whcn they are book. 1 have discussed elsewhere ( 3 ) cies c o ~ t l d make use oi this kind ot
rclatcti ro proof, lrut proot OK d ~ s p ~ o o f the value of Fermr's notelnook method, touchstone It is not true that all sci-
ctrmcs first ,tnd 1s in fact sttongest the effect ~t hat1 on his colleagues and ence is equal, or that we c a n n ~ l tjustly
&hen rt is absolutely convincing with- students, and the test~nlony that it compare the eRectiveness of scientist\
out any qua~ltitativemeasurement ' " c m be adopted by anyone with by any method other than J mutual-
O r l o sdy it another way, you can profit." recommendat~on system. The man to
c'~lch pherio~nena in a logical box o r It 1s true that ~t tahes great courtesy 1-ralch, the rnan to put your nroney on,
In d rric~lheitiatical hox. I h e logical to teach the method, especially to oi~e's is riot the man who wants to make
box 1s coarsc but strong The mathe- peers- o r thcir students. Thc strong- "a survey" or 1' "morc ~letailetf study"
m a t ~ c a l boy 1s fine-grained but fl~msy. rnference point of view is \o resolutely hut the n1~11i with thc notebook, the
7 he m'tthematical box is a beautiful critical o t methods of work and values nian with the alternat~vchypotheses and
way o l wrapping u p n problem, but in science that any attempt to com- the crucial exper~nlents, the man who
it wlll riot hold the p h e n o ~ n e n aunless p'rle specific cases IS l ~ k e l y to sound knows how to 'Inswer your Q~teation
they have been caught in a logical box both $mug and dustructive. Mainly one of disproof and is alreddy workiilg on
to b e g ~ nwith slro~tldt r j ta teach it by example and lt.
CVhat I arii saying 1s thnt, ~ r inumer- by e x h o r t ~ n gto self-analysis and selt- 'There are so111e really hard prob-
ous arcas that we call science, we Improvement only in general terms, as leniq, sorne h~gh-information problems,
have come to l ~ k eour habitual ways, I am d o ~ n ghere. ahead of us In reveral fields, problerris
,tnd our studles that can be continued But I w1!1 mentlon one severe but of photosynthesis, of cellular organita-
~ntlefin~tely We tiieasure, we define, useful private te\t -a touchstone of tion, of the n ~ o l c e u l astructure
~ and or-
we compulc, we analyze. but we d o strong inference-- that removes the nec- g a n i ~ a t i o n ot the ncrbous system. nor
not euclutie And this is not the n a y cssity f o r th~rd-person c r l t ~ c i s n ~be- , to mention sonic 01 our social and in
to use our minds rnost efiect~vely or cause lt is a test that anyone can learn ternational problems. It seems to tne
to mnke ihc fastest progress 11-1 solvlng to carry w ~ t hhim for use as necded. that the niethod of most rapid prog-
scientific cluest~ons. It is our old t r ~ e n dthe Racon~an "ex- ress in such complex areas, the most
Of course ~t Ir easy-and 1' 11 too clusion," but I call it "Thc Ques- efTective way of ~ising our brains. is
n - one xientist to call the t ~ o n" Obviously it should be applied
c o ~ ~ i n ~ o for going to be to set down expl~citly at
others unscientific hly point 1s not as n1~1cl.1to one's own thinking as to each step juat what the question is, and
th'lt my particular conclusions here are others' it consists of asking In your what all the alternatives are. and then
necessar~ly correct, but that we have own n ~ ~ n don , hearing any scientific to set up crucial experiments to try to
long needed sornc absolute standard o f explanation or theory put forward, "But disprove some. Prol~lemsof this c a n -
posslble scientific effectiveness by which bir, what experitixnt could disprove plexity, if they can be solved at al!,
to mea\ure how well we are succeed- your hypothesis?"; or, on hearing a scl- cdn be solved only by men generating
rng in varlous areas-a 5tandard that ctitific experiment described, "But sir, and excluding possibilities with niaxi-
many c o ~ l l d agree o n and one that what hypothesis does your experiment mum effectiveness, to obtain a high cle-
would be undistorted by the scientific cl~,prove?" gree of intornlation per unit time-nien
pre\sules and lashions of the tiines This goes straight t o the heart of w~llingto work a little bit at thinking.
ancf the vested ~nterests and busywork the nlatter. It forces everyone to re- When whole group.; of us begin ttr
that they develop. It is not public evalu- focus on the central question of whether concentrate like that, 1 believe we may
ation i a m ~nterestedin so much a\ a there is o r is not a testable scieiitific see the molecular-biology phenomenon
private measure by which to coinpare step torward. repeated over and over again, with or-
one's own scientific pertornlance with If such a questioil were asked aloud, der-of-magnitude increases in the rate
w h i t it m ~ g h tbe. 1 believe that strong man)., a s~ipposeclly great scientist of scientific understanding in almost
inference provicles this kind of stand- would sputter and turn l ~ v i dand would every field.
ard of what the maxitnuai possible scl- want to throw the questioner out, as
Refereaces and Notes
cntific efTectiveness could be-as well a hostile witness! Such a man is less
as a recipe for reaching it. than he appears, tor he is; obviously 1. A. \.I. Weinberg. iMitrerx~(r 1963, 159 (vvintcr
1963); Phys. T o d a y 17, 42 (1964).
not accustomed to think in terms of 2 . G. Po1)'3, Adathe?~~utics o?fd P ~ ~ l ~ , s i hRi fI!O S O I I -
alternative hypotheses and crucial ex- i r ? ~(Princeton IJniv. Presr, Princeton, N.J.,
1954), vol. 1, Irtdlrction and A ~ i n t o g y ill
Aids to Strong Inference perirnents for hirnself; and one might M o f l z e n ~ a f i c s ;vol. 2 , Pafterrls of Plnlr.rible
also wonder about the state of science lr~frrelrce
3. 5 . R. Platt, T h e k t ' c r t e ~ ~ i r fo~ff Scietilr
How can we learn the method and in the field h e is in. Rut who knows?- (Hoaghton MitTiiil. Boston. 1962); Eee espe-
c~ally chapters 4 and 8
teach it? It- is not difficult. 'The most the question might educate hini, and his 4 J D. Watson and F N C Crick, Natrfrr
important thing is to keep in mind that field too! 171, 707 (1953).
5. M. Meselsot~and F. Stahl, Proc. N a t l . Acod
this kind of thinking i s not a lucky On the other. hand, I think that Sci. U . S . 44, 671 (1958).
knack but a system that c a n be taught throughout most of molecular biology 6 . A. Rich, in Bioy71)sicnl Scirrice: A Sfrrdp

S C I E N C E , VOL. 146
Prograrn, J . L. Oncley et al., Eds. (Wiley, 13. F . Bacon, The Neiv Organon and Related of Minnesota for correspadence o n this
New York, 1959), p. 191. Writings (Liberal Arts Press, New York, article and a classroom reprint of it.
7. S. Benzer, Proc. Natl. Acad. Sci. U.S. 45, 1960), especially pp. 98, 112, 151, 156, 196. 16. M. Faraday, Faraday's Diary 1 8 2 0 4 2 (Bell,
1607 (1959). 14. K. R. Popper, Tlie Logic o f Scientific Dis- London, 1932-36).
8. J. Lederberg, Science 129, 1649 (1959). co1,er.v (Basic Books, York, 1959), p. 41. 17. H. L. Anderson and S. K. Allison, Rev.
9. P. F . Davison, D. Freifelder, B. W. Hollo- A modified view is given by T. S. Kuhn, The 27, 273 (1955).
M o d . P/IJJS.
way, J. Mol. Biol. 8, 1 (1964). Structirre o f Scierltific Rel~olutions (Univ, of 18. E. C . Watson [ A m . J. Phys. 13, 281 (1945)l
10. R. L. Garwin, L. M. Lederman, M. Wein- Chicago Press, Chicago, 1962), p. 146; it gives an English translation of both of
rich, Phys. Rex. 105, 1415 (1957). does not, I believe, invalidate any of these Roentgen's first papers on x-rays.
11. W. A. H . Rushton, personal communication. conclusions. 19. See G. W . Wheland, Adi,anced Organic
12. See G. F . Chew, M. Gell-Mann, A. H. Rosen- 15. T. C . Charnberlin, J . Geol. 5, 837 (1.897). I Chemistry (Wiley, New York, 1949), chapter
feld, Sci. A m . 210, 74 (Feb. 1964); ibid. 210, am indebted to Professors Preston Cloud 4, for numerous such examples.
60 (Apr. 1964); ibid. 210, 54 (June 1964). and Bryce Crawford, Jr., of the University 20. B. K. Forscher, Science 142, 339 (1963).

standing of deep and fundamental earth


processes that will probably forever be
hidden from direct investigation.
This article sketches recent ideas and
Glacier Geophysics measurements bearing on the glacier
flow process, its expression in dynamic
response of glaciers, and the possibility
Dynamic response of glaciers to changing climate of using these concepts to increase our
understanding of solid deformation
may shed light on processes in the earth's interior. processes in the earth generally.

Barclay Kamb
Glacier Flow

Flow velocities of most valley gla-


ciers are in the range of 0.1 to 2 meters
I n ordinary experience ice is a strong, monitors of climate. T o provide such per day, and are in general an increas-
rigid substance, and to casual observa- a connection requires an intimate ing function of glacier size and valley
tion glaciers appear to be solidly per- understanding of the dynamics of gla- slope. I n ice falls, like those seen in the
manent features of the landscapes cier motion. It has only recently been cover photograph, velocities of up to
where they occur. The finding that these realized, for example, that there are in- about 6 meters per day occur com-
great masses of ice are actually in mo- herent instabilities in glacier response, monly. Exceptional motions of as much
tion, flowing like fluids, has long at- so that a small climatic change may as 30 meters per day have been reported
tracted attention and has provoked produce a large glacial advance or re- for the great glaciers flowing from
much scientific controversy ( 1 ) . Only treat. the ice sheet into fjords on the west
within the past 15 years or so have we The flow of glaciers also serves as a coast of Greenland, such as Jakobs-
arrived at an understanding of how reminder that, on a time scale of mil- havn.
and why glaciers flow, through concepts lions of years, "solid" rocks themselves These velocities are actual downslope
of solid-state physics and of the new flow like fluids, in the conlplex and ice motions under gravity. The apparent
materials sciences and through devel- little understood processes by which the advancing or retreating movement of
opment of new and better means of earth builds great mountain ranges and the glacier snout (terminus) represents,
making physical measurements on other structures of continental dimen- of course, a balance between ice wast-
glaciers. sions, the processes of tectonophysics. age (ablation) and forward motion,
Ten percent of the earth's land area In this siniilarity, glacier geophysics and and is usually much slower than the ice
is at present covered by ice; during the tectonophysics have an important meet- motion itself. A striking exception is the
recent ice ages the ice-covered area was ing ground. Glaciers constitute great movement of certain glaciers, such as
almost three times as large. On at least outdoor laboratories in which concepts the famous Black Rapids Glacier of
four occasions ice invaded most of Can- and theories derived fro111 indoor lab- Alaska, which at times advance sud-
ada, the northern part of the United oratory experimentation can be tested denly and catastrophically at speeds of
Stales, northern Europe, and many on time scales and distance scales more up to 5 0 o r even 100 meters per day
mountain regions of the world. There nearly appropriate to the phenomena of (3).
is no general agreement yet 011 the solid-earth deformation, and yet still Although velocity measurements are
cause of this phenomenon ( 2 ) . In the accessible to human observation and made by long-established methods of
search for an explanation, glacier geo- measurement. Flow in glaciers produces precise surveying, only recently have
physics aims to provide a quantitative striking internal structures (see cover) the motions of individual glaciers been
connection between climatic change and that are analogs of structures in certain measured in sufficient detail to define
glacier fluctuation, so that glaciers can nletanlorphic rocks from which great completely the velocity field at the sur-
be interpreted confidently as long-term deformations of the earth's crust have face of the ice. Data for a portion of
been inferred. By studying how and Saskatchewan Glacier (Canada), one
The author is profeqsor of geology and geo- why these structures originate in gla- of the most completely studied so far
physics a t the California Institute of Technology,
Pasadena. ciers we can hope to get a better under- ( 4 ) , are shown in Fig. 1. Figure 16
16 OCTOBER 1964 353

Das könnte Ihnen auch gefallen