FROM pages 3-15 of:

Logical Foundations of Probability (1950).

by RUDOLF CARNAP
[p. 3:]
2. On the C!"#$#%!t#&n &$ !n '(p#%!n)*+
By the procedure of explication we mean the transformation of an inexact
prescientific concept the explicandum into a new exact concept the explicatum!
"lthough the explicandum cannot #e gi$en in exact terms it should #e made as clear as
possi#le #y informal explanations and examples!
%he tas& of explication consists in transforming a gi$en more or less inexact concept
into an exact one or rather in replacing the first #y the second! 'e call the gi$en
concept (or the term used for it) the explicandum and the exact concept proposed to
ta&e the place of the first (or the term proposed for it) the explicatum! %he explicandum
may #elong to e$eryday language or to a pre$ious stage in the de$elopment of scientific
language! %he explicatum must #e gi$en #y explicit rules for its use for example #y a
definition which incorporates it into a well-constructed system of scientific either
logicomathematical or empirical concepts!
%he term *explicatum* has #een suggested #y the following two usages! +ant calls a
,udgment explicati$e if the predicate is o#tained #y analysis of the su#,ect! -usserl in
spea&ing a#out the synthesis of identification #etween a confused nonarticulated sense
and a su#se.uently intended distinct articulated sense calls the latter the */xpli&at* of
the former! (For #oth uses see 0ictionary Of 1hilosophy 213456 ed! 0! Runes p! 175)!
'hat 8 mean #y *explicandum* and *explicatum* is to some extent similar to what 9! -!
:angford calls *analysandum* and *analysans*: ;the analysis then states an appropriate
relation of e.ui$alence #etween the analysandum and the analysans; (;%he notion of
analysis in Moore*s philosophy; in %he 1hilosophy of <! /! Moore 213436 ed! 1! "!
=chilpp pp! 351-45> see p! 353)> he says that the moti$e of an analysis ;is usually that
of supplanting a relati$ely $ague idea #y a more precise one; (i#id! p! 353)!
(1erhaps the form *explicans* might #e considered instead of *explicatum*> howe$er 8
thin& that the analogy with the terms *definiendum* and *definiens* would not #e useful
#ecause if the explication consists in gi$ing an explicit definition then #oth the
definiens and the definiendum in this definition express the explicatum while the
explicandum does not occur!) %he procedure of explication is here understood in a wider
sense than the procedures of analysis and clarification which +ant -usserl and
:angford ha$e in mind! %he explicatum (in my sense) is in many cases the result of an
analysis of the explicandum (and this has moti$ated my choice of the terms)> in other
cases howe$er it de$iates deli#erately from the explicandum #ut still ta&es its place in
some way> this will #ecome clear #y the su#se.uent examples!
" pro#lem of explication is characteristically different from ordinary scientific
(logical or empirical) pro#lems where #oth the datum and the solution are under
fa$ora#le conditions formulated in exact terms (for example *'hat is the product of 3
and 5?* *'hat happens when an electric current goes through water?*)! 8n a pro#lem of
explication the datum [p. ,:] $i@! the explicandum is not gi$en in exact terms> if it
were no explication would #e necessary! =ince the datum is inexact the pro#lem itself
is not stated in exact terms> and yet we are as&ed to gi$e an exact solution! %his is one
of the pu@@ling peculiarities of explication! 8t follows that if a solution for a pro#lem of
explication is proposed we cannot decide in an exact way whether it is right or wrong!
=trictly spea&ing the .uestion whether the solution is right or wrong ma&es no good
sense #ecause there is no clear-cut answer! %he .uestion should rather #e whether the
proposed solution is satisfactory whether it is more satisfactory than another one and
the li&e! 'hat is meant #y these .uestions will soon #e made clearer!
Before we turn to the chief .uestion $i@! what are the re.uirements for a satisfactory
solution of a pro#lem of explication that is to say for a satisfactory explicatum let us
loo& somewhat more at the way in which the pro#lem is to #e stated that is how the
explicandum is to #e gi$en! %here is a temptation to thin& that since the explicandum
cannot #e gi$en in exact terms anyway it does not matter much how we formulate the
pro#lem! But this would #e .uite wrong! On the contrary since e$en in the #est case we
cannot reach full exactness we must in order to pre$ent the discussion of the pro#lem
from #ecoming entirely futile do all we can to ma&e at least practically clear what is
meant as the explicandum! 'hat X means #y a certain term in contexts of a certain &ind
is at least practically clear to Y if Y is a#le to predict correctly X*s interpretation for most
of the simple ordinary cases of the use of the term in those contexts! 8t seems to me
that in raising pro#lems of analysis or explication philosophers $ery fre.uently $iolate
this re.uirement! %hey as& .uestions li&e: *'hat is causality?* *'hat is life?* *'hat is
mind?* *'hat is ,ustice?* etc! %hen they often immediately start to loo& for an answer
without first examining the tacit assumption that the terms of the .uestion are at least
practically clear enough to ser$e as a #asis for an in$estigation for an analysis or
explication! /$en though the terms in .uestion are unsystematic inexact terms there are
means for reaching a relati$ely good mutual understanding as to their intended meaning!
"n indication of the meaning with the help of some examples for its intended use and
other examples for uses not now intended can help the understanding! "n informal
explanation in general terms may #e added! "ll explanations of this &ind ser$e only to
ma&e clear what is meant as the explicandum> they do not yet supply an explication
say a definition of the explicatum> they #elong still to the formulation of the pro#lem
not yet to the construction of an answer! (/xamples! 1! 8 might say for example: ;8
mean #y the explicandum *salt* not its wide sense which it has in chemistry #ut its
[p. 5:] narrow sense in which it is used in the household language;! %his explanation is
not yet an explication> the latter may #e gi$en for instance #y the compound
expression *sodium chloride* or the synonymous sym#ol *Aa9l* of the language of
chemistry! 5! ;8 am loo&ing for an explication of the term *true* not as used in phrases
li&e *a true democracy* *a true friend* etc! #ut as used in e$eryday life in legal
proceedings in logic and in science in a#out the sense of *correct* *accurate* *$eridical*
*not false* *neither error nor lie* as applied to statements assertions reports stories
etc!; %his explanation is not yet an explication> an explication may #e gi$en #y a
definition within the framewor& of semantical concepts for example #y %ars&i*s
definition of *true* in 2'ahrheits#egriff6 (for a##re$iated titles in s.uare #rac&ets see the
Bi#liography at the end of this $olume) or #y 01B-1 #elow! By explanations of this
&ind the reader may o#tain step #y step a clearer picture of what is intended to #e
included and what is intended to #e excluded> thus he may reach an understanding of
the meaning intended which is far from perfect theoretically #ut may #e sufficient for
the practical purposes of a discussion of possi#le explications!
3. Re-*#"e+ent. $&" !n '(p#%!t*+
" concept must fulfil the following re.uirements in order to #e an ade.uate
explicatum for a gi$en explicandum: (1) similarity to the explicandum (5) exactness
(3) fruitfulness (4) simplicity!
=uppose we wish to explicate a certain prescientific concept which has #een
sufficiently clarified #y examples and explanations as ,ust discussed! 'hat is the
explication of this concept intended to achie$e? %o say that the gi$en prescientific
concept is to #e transformed into an exact one means of course that an exact concept
corresponding to the gi$en concept is to #e introduced! 'hat &ind of correspondence is
re.uired here #etween the first concept the explicandum and the second the
explicatum?
=ince the explicandum is more or less $ague and certainly more so than the
explicatum it is o#$ious that we cannot re.uire the correspondence #etween the two
concepts to #e a complete coincidence! But one might perhaps thin& that the explicatum
should #e as close to or as similar with the explicandum as the latter*s $agueness
permits! -owe$er it is easily seen that this re.uirement would #e too strong that the
actual procedure of scientists is often not in agreement with it and for good reasons! :et
us consider as an example the prescientific term *fish*! 8n the construction of a
systematic language of @oClogy the concept Fish designated #y this term has #een
replaced #y a scientific concept designated #y the same term [p. /:] *fish*> let us use for
the latter concept the term *piscis* in order to a$oid confusion! 'hen we compare the
explicandum Fish with the explicatum 1iscis we see that they do not e$en
approximately coincide! %he latter is much narrower than the former> many &inds of
animals which were su#sumed under the concept Fish for instance whales and seals
are excluded from the concept 1iscis! 2%he situation is not ade.uately descri#ed #y the
statement: *%he pre$ious #elief that whales (in <erman e$en called *'alfische*) are also
fish is refuted #y @oClogy*! %he prescientific term *fish* was meant in a#out the sense of
*animal li$ing in water*> therefore its application to whales etc! was entirely correct!
%he change which @oClogists #rought a#out in this point was not a correction in the field
of factual &nowledge #ut a change in the rules of the language> this change it is true
was moti$ated #y factual disco$eries!6 %hat the explicandum Fish has #een replaced #y
the explicatum 1iscis does not mean that the former term can always #e replaced #y the
latter> #ecause of the difference in meaning ,ust mentioned this is o#$iously not the
case! %he former concept has #een succeeded #y the latter in this sense: the former is no
longer necessary in scientific tal&> most of what pre$iously was said with the former can
now #e said with the help of the latter (though often in a different form not #y simple
replacement)! 8t is important to recogni@e #oth the con$entional and the factual
components in the procedure of the @oClogists! %he con$entional component consists in
the fact that they could ha$e proceeded in a different way! 8nstead of the concept 1iscis
they could ha$e chosen another conceptDlet us use for it the term *piscisE* D which
would li&ewise #e exactly defined #ut which would #e much more similar to the
prescientific concept Fish #y not excluding whales seals etc! 'hat was their moti$e for
not e$en considering a wider concept li&e 1iscisE and instead artificially constructing
the new concept 1iscis far remote from any concept in the prescientific language? %he
reason was that they reali@ed the fact that the concept 1iscis promised to #e much more
fruitful than any concept more similar to Fish! " scientific concept is the more fruitful
the more it can #e #rought into connection with other concepts on the #asis of o#ser$ed
facts> in other words the more it can #e used for the formulation of laws! %he @oClogists
found that the animals to which the concept Fish applies that is those li$ing in water
ha$e #y far not as many other properties in common as the animals which li$e in water
are cold-#looded $erte#rates and ha$e gills throughout life! -ence the concept 1iscis
defined #y these latter properties allows more general statements than any concept
defined so as to #e more similar to Fish> and this is what ma&es the concept 1iscis more
fruitful! [p. 0:]
8n addition to fruitfulness scientists appreciate simplicity in their concepts! %he
simplicity of a concept may #e measured in the first place #y the simplicity of the form
of its definition and second #y the simplicity of the forms of the laws connecting it
with other concepts! %his property howe$er is only of secondary importance! Many
complicated concepts are introduced #y scientists and turn out to #e $ery useful! 8n
general simplicity comes into consideration only in a case where there is a .uestion of
choice among se$eral concepts which achie$e a#out the same and seem to #e e.ually
fruitful> if these concepts show a mar&ed difference in the degree of simplicity the
scientist will as a rule prefer the simplest of them!
"ccording to these considerations the tas& of explication may #e characteri@ed as
follows! 8f a concept is gi$en as explicandum the tas& consists in finding another
concept as its explicatum which fulfils the following re.uirements to a sufficient degree!
1! %he explicatum is to #e similar to the explicandum in such a way that in most
cases in which the explicandum has so far #een used the explicatum can #e used>
howe$er close similarity is not re.uired and considera#le differences are permitted!
5! %he characteri@ation of the explicatum that is the rules of its use (for instance in
the form of a definition) is to #e gi$en in an exact form so as to introduce the
explicatum into a well-connected system of scientific concepts!
3! %he explicatum is to #e a fruitful concept that is useful for the formulation of
many uni$ersal statements (empirical laws in the case of a nonlogical concept logical
theorems in the case of a logical concept)!
4! %he explicatum should #e as simple as possi#le> this means as simple as the more
important re.uirements (1) (5) and (3) permits!
1hilosophers scientists and mathematicians ma&e explications $ery fre.uently! But
they do not often discuss explicitly the general rules which they follow implicitly! "
good explicit formulation is gi$en #y +arl Menger in connection with his explication of
the concept of dimension (;'hat is dimension?; "mer! Math! Monthly 57 213436 5-B>
see p! 5: F 3 ;9riteria for a satisfactory definition; 2explication in our terminology6)! -e
states the following re.uirements! %he explicatum ;must include all entities which are
always denoted and must exclude all entities which are ne$er denoted; #y the
explicandum! %he explication ;should extend the use of the word #y dealing with
o#,ects not &nown or not dealt with in ordinary language! 'ith regard to such entities a
definition 2explication6 cannot help #eing ar#itrary!; %he explication ;must yield many
conse.uences; theorems possessing ;generality and simplicity; and connecting the
explicatum with concepts of other theories! =ee also the discussions #y 9! -! :angford
referred to in F 5!
Terminological remarks. 1! %he word *concept* is used in this #oo& as a [p. 1:]
con$enient common designation for properties relations and functions! 2Aote that (a) it
does not refer to terms i!e! words or phrases #ut to their meanings and (#) it does not
refer to mental occurrences of concei$ing #ut to something o#,ecti$e!6 For more
detailed explanations see 2=emantics6 p! 537> 2Meaning6 p! 51! 5! 8f 8 spea& a#out an
expression (e!g! a word a phrase a sentence etc!) in distinction to what is meant or
designated #y it 8 include it in .uotation mar&s! %hat this distinction is necessary in
order to a$oid confusion has #ecome more and more clear in the recent de$elopment of
logic and analysis of language! 3! 8f 8 want to spea& a#out a concept (property relation
or function) designated #y a word 8 sometimes use the de$ice of capitali@ing the word
especially if it is not a noun (compare 2Meaning6 p! 1B n!)! For example 8 might write
*the relation 'armer*> to write instead *the relation warmer* would too& strange and #e
contrary to /nglish grammar> to write *the relation of x #eing warmer than y* would #e
incon$enient #ecause of its length> the customary way of writing *the relation *warmer**
would not #e .uite correct #ecause *warmer* is not a relation #ut a word designating a
relation! =imilarly 8 shall sometimes write: *the property (or concept) Fish* (instead of
*the property of #eing a fish*)> *the property (or concept) Red* (instead of *the property of
#eing red* or *the property of redness*) and the li&e!
"rne Aaess defines and uses a concept which seems related to our concept
/xplicatum (;8nterpretation and preciseness! 8! =ur$ey of #asic concepts; 2Oslo
Gni$ersitetets =tudent&ontor 134B6 2mimeographed6> this is the first chapter of a
forthcoming #oo&)! Aaess defines *the formulation G is more precise than % (in the
sense that G may with profit #e su#stituted for %)* #y *there are interpretations of %
which are not interpretations of G #ut there are no interpretations of G which are not
also interpretations of %* (ibid! p! 3H)! %his comparati$e concept ena#les Aaess to deal
with a series of consecuti$e ;precisations; of a gi$en concept! Aaess announces that a
later chapter (iii) of the #oo& will #e ;de$oted to the .uestion of how to measure
degrees of am#iguity $agueness and similar properties;! %he comparati$e concept
mentioned and these .uantitati$e concepts may pro$e to #e effecti$e tools for a more
penetrating analysis of explication!
,. C!..#$#%!t&"y2 C&+p!"!t#3e2 !n) 4*!nt#t!t#3e C&n%ept.
" classificatory concept (e!g! 'arm) ser$es for classifying things into two &inds! "
comparati$e concept is a relation #ased on a comparison with the sense of *more (in a
certain respect)* (e!g! 'armer) or *more or e.ual*! " .uantitati$e concept ser$es to
descri#e something with the help of numerical $alues (e!g! temperature)!
"mong the &inds of concept used in science three are of special importance! 'e call
them classificatory comparati$e and .uantitati$e concepts! 'e shall ma&e use of this
distinction in our later discussion of confirmation and pro#a#ility! 8n prescientific
thin&ing classificatory [p. 9:] concepts are used most fre.uently! 8n the course of the
de$elopment of science they are replaced in scientific formulations more and more #y
concepts of the two other &inds although they remain always useful for the formulation
of o#ser$ational results! Classificatory concepts are those which ser$e for the
classification of things or cases into two or a few mutually exclusi$e &inds! %hey are
used for example when su#stances are di$ided into metals and nonmetals and again
the metals into iron copper sil$er etc!> li&ewise when animals and plants are di$ided
into classes and further di$ided into orders families genera and finally species> when
the things surrounding us are descri#ed as warm or cold #ig or small hard or soft etc!
or when they are classified as houses stones ta#les men etc! 8n these examples the
classificatory concepts are properties! 8n other cases they are relations for example
those designated #y the phrases *x is close to y* and *the person x is ac.uainted with the
field of science y*! (" relation may #e regarded as a property of ordered pairs!)
Quantitative concepts (also called metrical or numerical concepts or numerical
functions) are those which ser$e for characteri@ing things or e$ents or certain of their
features #y the ascription of numerical $alues> these $alues are found either directly #y
measurement or indirectly #y calculation from other $alues of the same or other
concepts! /xamples of .uantitati$e concepts are length length of time $elocity
$olume mass force temperature electric charge price 8!I! infantile mortality etc! 8n
many cases a .uantitati$e concept corresponds to a classificatory concept! %hus
temperature corresponds to the property 'arm> and the concept of a distance of less
than fi$e miles corresponds to the relation of proximity! %he method of .uantitati$e
concepts and hence of measurement was first used only for physical e$ents #ut later
more and more in other fields also especially in economics and psychology!
Iuantitati$e concepts are no dou#t the most effecti$e instruments in the scientific
arsenal! =ometimes scientists especially in the fields of social science and psychology
hold the $iew that in cases where no way is disco$ered for the introduction of a
.uantitati$e concept nothing remains #ut to use concepts of the simplest &ind that is
classificatory ones! -ere howe$er they o$erloo& the possi#ility and usefulness of
comparative concepts, which in a sense stand #etween the two other &inds!
9omparati$e concepts (sometimes called topological or order concepts) ser$e for the
formulation of the result of a comparison in the form of a more-less-statement without
the use of numerical $alues! Before the scientific .uantitati$e concept of temperature
was introduced e$eryday language contained comparati$e concepts! 8nstead of merely
classifying things into a few &inds [p. 10:] with the help of terms li&e *hot* *warm*
*lu&e-warm* *cold* a more effecti$e characteri@ation was possi#le #y saying that x is
warmer than y (or colder or e.ually warm as the case may #e)!
" comparati$e concept is always a relation! 8f the underlying classificatory concept is
a property (e!g! 'arm) the comparati$e concept is a dyadic relation that is one with
two arguments (e!g! 'armer)! 8f the classificatory concept is a dyadic relation (e!g! the
relation of x #eing ac.uainted with (the field) y) the comparati$e concept has in
general four arguments (e!g! the relation of x #eing #etter ac.uainted with y than u with
v)! 8t is sometimes useful to regard the tetradic relation as a dyadic relation #etween two
pairs! ('e might say for example: *the relation of #eing ac.uainted holds for the pair x,
y to a higher degree than for the pair u, v *!) =ometimes the introduction of a triadic
relation is preferred to that of a tetradic relation! 8f we do not &now how to compare the
degree of 1eter*s &nowledge in physics with Jac&*s &nowledge in history we might
perhaps #e content to use either or #oth of the two triadic relations expressed #y the
following phrases: *x is #etter ac.uainted with (the field) y than with v * * x is #etter
ac.uainted with y than u *! %he first of these two relations re.uires that we are a#le to
compare the degree of 1eter*s &nowledge in physics with that in history which might
seem pro#lematical! %he second relation in$ol$es the comparison of 1eter*s &nowledge
in physics with that of Jac&> here it seems easier to in$ent suita#le tests!
/ach of the comparati$e concepts gi$en a#o$e as an example has the meaning of
*more* or *to a higher degree* with respect to a gi$en classificatory concept! %o any of
those classificatory concepts (e!g! 'arm) we can li&ewise construct a comparati$e
concept meaning *less* or *to a lower degree* (e!g! :ess-warm> in other words 9older)>
this is the con$erse of the first comparati$e concept! 8n either case the comparati$e
concept regarded as a dyadic relation (of simple entities pairs etc!) has o#$iously the
following relational properties: it is irreflexi$e transiti$e and (hence) asymmetric! (For
definitions of these and other terms of the theory of relations see 055-5!) Aote!
8n addition to the form of comparati$e concepts ,ust mentioned there is another form
less customary #ut often more useful! " concept of this second &ind does not mean
*more* #ut *more or e.ual* with respect to the underlying classificatory concept in other
words *to at least the same degree* that is *to the same or a higher degree* (e!g! the
relation of x #eing at least as warm as y)! Or it may mean *less or e.ual* (e!g! the
relation of x #eing less warm than y or e.ually as warm as y> in other words of x #eing
at most as warm as y)! 8t is easily seen that a comparati$e [p. 11:] concept of this second
&ind regarded as a dyadic relation is reflexi$e and transiti$e #ut neither symmetric nor
asymmetric! " comparati$e relation is sometimes of such a &ind that for any x and y, it
holds either #etween x and y or #etween y and x (or #oth)! 8n this case the relation (for
example 'armer-Or-/.ually-'arm) orders its mem#ers in a &ind of linear order! 8f
howe$er the condition is not fulfilled then there are incompara#le cases! %hus it might
perhaps #e that we find it possi#le to compare the scientific achie$ements of two
persons if #oth wor& in the same field while we do not &now a way of comparing a
physicist with a historian!
8n e$eryday language the first form of comparati$e concept is much more customary
than the second! %here are many single words for those of the first form for instance
*a#o$e* *#eyond* *after* etc! and especially the comparati$es for instance *more*
*warmer* etc! while there are hardly any single words for those of the second form! On
the other hand there is a general trend in the de$elopment of the language of science
toward concepts which are wider than corresponding concepts of prescientific language
#y including extreme cases especially cases of @ero $alue or of identity or e.uality> for
example the term *num#er* is now ta&en as including 7 *class* as including the null
class *$elocity* as including the case of rest regarded as $elocity 7 etc! 'ith respect to
comparati$e concepts this trend means a de$elopment from those of the first &ind to
those of the second #ecause the latter include the #oundary case of e.uality! One
ad$antage of those of the second &ind consists in the fact that on the #asis of *more or
e.ual* we can define #oth *e.ual* and *more* (*x K y* can #e defined #y *x y !n) y x*> *x
L y* #y *x y !n) n&t y x*) while on the #asis of *more* we cannot define either *e.ual*
or *more or e.ual*! For these reasons when we come to a discussion of a comparati$e
concept of confirmation (F H) we shall ta&e one of the second form as expressed #y: *&
is confirmed #y e to the same or a higher degree than &M #y eM *!
For an analysis of comparati$e and .uantitati$e concepts and an explanation of the
steps to #e ta&en in the construction of concepts of these &inds see 9arnap
1hysi&alische Begriffs#ildung (+arlsruhe 135N)! 9! <! -empel and 1! Oppenheim ha$e
de$eloped and impro$ed the characteri@ations of the two &inds of concept and
illustrated their roles in $arious fields of science in their #oo& 0er %ypus#egriff im
:ic&te der neuen :ogi&: 'issenschaftstheoretische Gntersuchungen sur
+onstitutionsforschung und 1sychologie (:eiden 133N)!
5. C&+p!"!t#3e !n) 4*!nt#t!t#3e C&n%ept. !. '(p#%!t!
%he role of comparati$e and .uantitati$e concepts as explicata is discussed in
preparation for a later discussion of comparati$e and .uantitati$e concepts of
confirmation! [p. 12:]
9lassificatory concepts are the simplest and least effecti$e &ind of concept!
9omparati$e concepts are more powerful and .uantitati$e concepts still more> that is to
say they ena#le us to gi$e a more precise description of a concrete situation and more
important to formulate more comprehensi$e general laws! %herefore the historical
de$elopment of the language is often as follows: a certain feature of e$ents o#ser$ed in
nature is first descri#ed with the help of a classificatory concept> later a comparati$e
concept is used instead of or in addition to the classificatory concept> and still later a
.uantitati$e concept is introduced! (%hese three stages of de$elopment do of course not
always occur in this temporal order!)
%he situation may #e illustrated with the help of the example of those concepts which
ha$e led to the .uantitati$e concept of temperature! %he state of #odies with respect to
heat can #e descri#ed in the simplest and crudest way with the help of classificatory
concepts li&e -ot 'arm and 9old (and perhaps a few more)! 'e may imagine an early
not recorded stage of the de$elopment of our language where only these classificatory
terms were a$aila#le! :ater an essential refinement of language too& place #y the
introduction of a comparati$e term li&e *warmer*! 8n the case of this example as in many
others this second step was already made in the prescientific language! Finally the
corresponding .uantitati$e concept that of temperature was introduced in the
construction of the scientific language!
%he concept %emperature may #e regarded as an explicatum for the comparati$e
concept 'armer! %he first of the re.uirements for explicata discussed in F3 that of
similarity or correspondence to the explicandum means in the present case the
following: %he concept %emperature is to #e such that in most cases if x is warmer than
y (in the prescientific sense #ased on the heat sensations of the s&in) then the
temperature of x is higher than that of y! -ere a few remar&s may #e made!
(i) %he re.uirement refers to most cases not to all cases! 8t is easily seen that the
re.uirement is fulfilled only in this restricted sense! =uppose 8 enter a moderately heated
room twice first coming from an o$erheated room and at a later time coming from the
cold outside! %hen it may happen that 8 declare the room on the #asis of my sensations
to #e warmer the second time than the first while the thermometer shows at the second
time the same temperature as at the first (or e$en a slightly higher one)! /xperiences of
this &ind do not at all lead us to the conclusion that the concept %emperature defined
with reference to the thermometer is inade.uate as an explicatum for the concept
'armer! On the contrary we ha$e #ecome accustomed to let the scientific concept
o$errule the prescientific [p. 13:] one in all cases of disagreement! 8n other words the
term *warmer* has undergone a change of meaning! 8ts meaning was originally #ased
directly on a comparison of heat sensations #ut after the acceptance of the scientific
concept %emperature into our e$eryday language the word *warmer* is used in the sense
of *ha$ing a higher temperature*! %hus the experience descri#ed a#o$e is now
formulated as follows: ;8 #elie$ed that the room was at the second time warmer than at
the first #ut this was an error> the room was actually not warmer> 8 found this out with
the help of the thermometer;! For this second scientific meaning of *warmer* we shall
use in the following discussion the term *warmerE*!
(ii) %he con$erse of the re.uirement mentioned a#o$e would #e this: the concept
%emperature is to #e such that if x is not warmer than y (in the prescientific sense) then
the temperature of x is not higher than that of y! 8t is important to reali@e that this is not
re.uired not e$en ;in most cases;! 'hen the difference #etween the temperatures of x
and y is small then as a rule we notice no difference in our heat sensations! %his again
is not ta&en as a reason for re,ecting the concept %emperature! On the contrary here
again we ha$e #ecome accustomed to the new scientific concept 'armerE and thus we
say: ;z is actually warmerE than y although we cannot feel the difference;!
(iii) %hus we ha$e two scientific concepts corresponding to the prescientific concept
'armer! %he one is the comparati$e concept 'armerE the other the .uantitati$e
concept %emperature! /ither of them may #e regarded as an explicatum of 'armer!
Both are defined with reference to the thermometer! =ince the thermometer has a higher
discriminating power than our heat sensations #oth scientific concepts are superior to
the prescientific one in allowing more precise descriptions! %he procedure leading from
the explicandum to either of the two explicata is as follows! "t first the prescientific
concept is guiding us in our choice of an explication (with possi#le exceptions as
discussed earlier)! Once an explicatum is defined in a relati$ely simple way we follow
its guidance in cases where the prescientific concept is not sufficiently discriminati$e! 8t
would #e possi#le #ut highly inad$isa#le to define a concept %emperature in such a way
that x and y are said to ha$e the same temperature whene$er our sensations do not show
a difference! %his concept would #e in closer agreement with the explicandum than the
concept %emperature actually used! But the latter has the ad$antage of much greater
simplicity #oth in its definition D in other words its method of measurement D and in
the laws formulated with its help!
(i$) Of the two scientific terms *warmerE* and *temperature* the latter [p. 1,:] is the
one important for science> the former ser$es merely as a con$enient a##re$iation for
*ha$ing a higher temperature*! %he .uantitati$e concept %emperature has pro$ed its great
fruitfulness #y the fact that it occurs in many important laws! %his is not always the case
with .uantitati$e concepts in science e$en if they are well defined #y exact rules of
measurement! For instance it has sometimes occurred in psychology that a .uantitati$e
concept was defined #y an exact description of tests #ut that the expectation of finding
laws connecting the $alues thus measured with $alues of other concepts was not
fulfilled> then the concept was finally discarded as not fruitful! 8f it is a .uestion of an
explication of a prescientific concept then a situation of the &ind descri#ed where we
do not succeed in finding an ade.uate .uantitati$e explicatum ought not to discourage
us altogether from trying an explication! 8t may #e possi#le to find an ade.uate
comparati$e explicatum! :et us show this #y a fictitious example! %he experience
leading to the concept %emperature was first a comparati$e one> it was found that if x is
warmer than y (in the prescientific sense) and we #ring a #ody of mercury first in
contact with x and later with y then it has at the first occasion a greater $olume than at
the second! By a certain de$ice it was made possi#le to measure the small differences in
the $olume of the mercury> and that was ta&en as #asis for the .uantitati$e concept
%emperature! Aow let us assume fictitiously that we did not find technical means for
measuring the differences in the $olume of the mercury although we were a#le to
o#ser$e whether the mercury expands or contracts! 8n this case we should ha$e no #asis
for a .uantitati$e concept %emperature #ut it would still #e possi#le to define the
comparati$e concept 'armerE with reference to an expansion of the! mercury! %his
scientific concept 'armerE could then #e ta&en as explicatum for the prescientific
concept 'armer! -ere in the fictitious case the concept 'armerE would #e of greater
importance than it is in actual physics #ecause it would #e the only explicatum! Aote
that 'armerE here is essentially the same concept as 'armerE in the earlier discussion
#ut that there is a difference in the form of the two definitions! 8n the former case we
defined 'armerE in terms of higher temperature hence with the help of a .uantitati$e
concept> here in the fictitious case it is defined with reference to the comparati$e
concept of the expansion of mercury without the use of .uantitati$e concepts! %he
distinction #etween these two ways of defining a comparati$e concept the .uantitati$e
way and the purely comparati$e that is non.uantitati$e way will #e of importance
later when we discuss the comparati$e concept of confirmation!
%o ma&e a wea&er fictitious assumption suppose that the $olume [p. 15:] differences
could #e measured and hence the .uantitati$e concept %emperature could #e defined #ut
that D this is the fictitious feature D no important laws containing this concept had
#een found! 8n this case the concept would #e discarded as not fruitful! "nd hence in
this case li&ewise the comparati$e concept 'armerE would #e ta&en as the only
explicatum for 'armer!
:ater when we discuss the pro#lem of explication for the concept of confirmation
we shall distinguish three concepts the classificatory the comparati$e and the
.uantitati$e concept of confirmation! %hey are analogous to the concepts 'arm
'armer and %emperature> thus the results of the present discussion will then #e
utili@ed!
2Boardman*s note: "lthough 8 am una#le to $erify these references (the :awrence
li#rary copy of Logical oundations of !robability ha$ing #een stolen) here are the
wor&s to which 9arnap pro#a#ly is referring in the excerpt:6
2'ahrheits#egriff6 K "! %ars&i ;0er 'ahrheits#egriff in den formalisierten =prachen;
1 "tudia !#ilosop#ica 5N1 (1335)!
2Meaning6 K R! 9arnap $eaning and %ecessity Gni$ersity of 9hicago1ress 134B!
2=emantics6 K R! 9arnap &ntroduction to "emantics -ar$ard Gni$ersity 1ress 1345