DIMENSIONS OF PSYCHOLOGICAL THEORY

RICHAKD W. COAN
University of Arizona

NE form of human behavior that has al- are accompanied by an emphasis on becoming,

O ways interested psychologists is the con-

struction of psychological theory. Possibly
because of our personal involvement in this be-
creativity, growth, and self-actualization and by a
conception of the learner as a concerned participant
in the learning process.
havior, however, (here has been little effort to sub- The dichotomotis classifications we have thus far
ject it to systematic scientific investigation. The considered are all reminiscent of one proposed by
main concern of the present paper is the basic Murray (1938) in the 1930s. Murray claimed
trends evident in psychological theory both over that there were two basic types of psychologists—-
time and at any given point in time. This seems the periphcralists and the centralists. The periph-
an appropriate site for multivariate research. Yet eral isls tend at the same time to be posit ivists,
with few exceptions, all the relevant analyses to mechanists, elemenfarists, sensationisfs, and ob-
date have been conducted on an armchair basis— jectivists, while the centralists tend to be con-
sometimes, fortunately, with rather penetrating ceptualists, totalists, intuitionists, dynamicists, and
insight. subjectivist.s. A still earlier and obvious parallel is
Every reader is already familiar with some of the distinction which James (1907) made between
these speculative analyses. Gordon Allport (19.SS) tender-minded and tough-minded philosophers.
has distinguished between a Lockcan tradition and The above schemes constitute a meager sampling
a Leibnitzean tradition in psychological theory. of pertinent speculation, but they represent a com-
The first has predominated in Anglo-American mon pattern that could easily be illustrated more
psychology. Its stress on a reactive organism may extensively. The classification is unidimensional,
be traced through associalionism to such modern and there is surprising agreement regarding the
manifestations as environmentalism, behaviorism, grouping of component variables—T say surprising
stimulus-response psychology, positivism and opera- because it is so easy to think of individual theorists
tionisrn, and a stress on the peripheral and molec- and movements that combine these variables into
ular. The Leibnitzean tradition assumes an active nonconforming patterns. There is much casual
or self-propelled organism, and its influence is best evidence to suggest that some of these variables
seen today on the European continent in such move- actually constitute mutually independent, or at
ments as phenomenology and Gestalt psychology. least semiindependent, dimensions.
In a similar vein, Rogers (1961) spoke of two Tn a somewhat more sophisticated treatment of
basic trends in present-day American psychology, this problem, Erunswik (19S2) insisted on what
which he called objective and existential. The amounts to a two-dimensional system. He dis-
first is characterized by rigorous hard-headedness, tinguished two basic issues in psychological theory.
reductionist theory, operational definitions, ob- The first is concerned with the rigor of fact finding,
jective methods, and an emphasis on the concrete inference, and communication, or with what we
and specific. The existential trend, on the other might call the quest for certainty. There are
hand, is concerned with the experiencing person and essentially two ways of handling this issue. One
with the whole spectrum of human behavior. is subjectivistic, mentalistic, and "introspection-
Ansbachcr (1961) employed a comparable di- istic"; the other is objectivistic. The first is as-
chotomy in a paper on holism. He recognized sociated with the rationalistic and idealistic out-
two basic viewpoints in psychology—the elemen- look prevalent on the European continent since
taristic and the holistic. Elementarism is said to be the time of Descartes. The second is associated
generally associated with determinism, mechanism, with (.he empiricism and positivism of England and
reductionism, and the espousal of a "spectator" the United States. The second issue is the
theory of knowledge, while holism and organicism level of complexity of theory. Here, of course,
715
716 AMERICAN PSYCHOLOGIST

theory ranges from the holistic to the elementaristic, as separate variables potentially applicable to different
or, in Tolman-Brunswik terms, from the molar to aspects of the same body of theory. The ultimate list com-
prised 34 variables gleaned from an earlier compilation
the molecular. of about twice this length. It is convenient to think of
According to Brunswik, there is a continuous his- Ihese as falling roughly into four categories, depending on
torical progression from subjectivism to objectivism whether they are concerned with content emphasis,
as well as from molecular to molar approaches. methodological emphasis, basic assumptions, or mode of
Within a sample of theorists well distributed over conceptualization. They include the following:
time, we might thus expect the two dimensions to ('ontent emphasis
be correlated, so that they might be less distinguish-
1. Learning
able empirically than they are logically. In the 2. Sensation and perception
light of Brunswik's more detailed analysis of his- 3. Motivation
torical trends, a total collapsing of the two di- 4. Conscious processes, conscious experience
mensions into a single factor seems unlikely, but 5. Observable behavior, action, performance
6. Unconscious processes
we should expect a positive association between
1. Emotion
objectivism and molarity. This expectation runs 8. Self-concept, self-perception
counter to the other speculation we have noted. It 9. Biological determinants of behavior
may well be that the trends that are seen are 10. Social determinants of behavior
governed by the theoretical biases of the viewer- - 11. Heredity, constitution
12. Influence of past experience on behavior
they depend, for example, on whether he favors a 13. Immediate external determinants of behavior
molar objectivism, as Brunswik does, or a molar 14. Total organization of behavior
subjectivism, as do most of the other psychologists 15. Uniqueness of individual personality
16. Persisting traits of individuals
mentioned.
PROCEDURE Methodological emphasis
17. Introspective reports of experience
It was hoped that an analysis of rated variables applied ] 8. Rigidly controlled experimentation
to a substantial number of psychological theorists would 19. Statistical analysis
reveal both historical trends and covariational trends 20. "Armchair" speculation
independent of time. The first task was to select an 21. Naturalistic observation
appropriate sample of theorists, and this was done on the
basis of the rated importance of their contributions to Bask assumptions
psychological theory.1 In the course of an investigation
reported earlier (Coan & Zagona, 1962), ratings were ob- 22. Voluntarism
tained for 142 theorists both with respect to overall im- 23. Determinism
portance and with respect to the significance of their con- 24. Finalism, teleology
tributions during specific decades in which they were 25. Mechanism
active, from the 1880s to the 1950s. In subsequent work,
attention has been confined to those theorists who emerged Mode of conceptualization
among the top SO in overall ratings or among the top 26. Operational definition of concepts
10 in the ratings for any decade. By virtue of the over- 27. Elementarism, atomism, description or analysis of
lapping of these lists, this constitutes a total of 54 different events in terms of relatively small units
theorists. 28. Holism, totalism, treatment of phenomena in rela-
The construction of a list of variables posed a different tively global terms
kind of problem. It was essential that the variables pro- 29. Nomolhetic approach, formulation of general prin-
vide a comprehensive coverage of all basically important ciples
aspects of psychological theory. At the same time, the 30. Normative generalization, statistical generalizations
list could not be too long if cooperation of raters was about groups of people
to be ensured. Furthermore, it was necessary that each 31. Quantitative formulation of principles and relation-
variable be applicable to all theorists and that it be fairly ships
clearly ratable. Words like "empirical" or "objective" that 32. Quantitative description of individuals and behavior
were unnecessarily vague or ambiguous had to be avoided. 33. Conceptualization in terms of hypothetical entities
In general, it seemed best to avoid any assumptions regard- 34. Use of analogies based on physical systems
ing inherent relationships among variables. For this reason,
for example, "holistic" and "elementaristic" were treated The next step, of course, was to relate the variables to
the theorists. Perhaps the ideal method would involve an
1
The author is grateful for the collaborative efforts of exhaustive content analysis of the published work of each
Salvatore V. Zagona in the early stages of this research. theorist. For the sake of securing a comprehensive picture
 DIMENSIONS OF PSYCHOLOGICAL THEORY 717

in a relatively brief period, it was decided instead that primary axes -were derived.2 Finally, a set of
ratings would be solicited from presumably qualified ex- factor scores was obtained for every theorist by
perts. For this purpose, a list of 232 correspondents was
multiple-regression procedures.
compiled. This contained an initial core of psychologists
widely dispersed throughout the United States who were We can most readily gain an impression of each
known either to have taught a course in the history of factor by noting the most salient variables to which
psychology or to have interests in the realm of history it contributes and by considering the theorists who
and systems. The core was supplemented by many ad- stand near the extremes of the continuum that it
ditional names of psychologists known through personal
represents. In the following lists two numerical
contact or through correspondence with others to have
appropriate interest and background. The theorists were values are reported for each variable; the first is a
divided into three equal subgroups, and each correspondent factor loading, and the second is the factor-variable
was asked initially to rate only 18 theorists, approximately correlation. Each list of theorists starts with the
one third of the correspondents receiving each list of 18. individual whose factor score is most extreme,
In a follow-up mailing, each correspondent was given
whether positive or negative.
an opportunity to rate the 36 remaining theorists. In
both instances, the rater was permitted to confine his
Factor 1
ratings to those theorists with whom he felt sufficiently
familiar, but he was requested to furnish a complete set 4. Conscious processes, 1.07, .92
of ratings for each theorist that he selected. The number 17. Introspective reports, .96, .96
of complete sets of ratings returned for any theorist varied 22. Voluntarism, .81, .92
from six (for Charcot) to 38 (for Freud), the average 24. Finalism, .76, .86
being about 20 sets per theorist. All rating was done on a 20. "Armchair" speculation, .70, .87
S-point scale on which +2 represented marked positive 2. Sensation and perception, .55, .31
emphasis on the given variable, +1 represented slight 6. Unconscious processes, .54, .66
positive emphasis, 0 represented no particular emphasis 8. Self-concept, .55, .76
either way, —1 represented slight rejection of the given 33. Hypothetical entities, .54, .26
variable, and —2 represented marked rejection. 5. Observable behavior, —.95, —.84
The present method of data collection is certainly 23. Determinism, —.60, —.82
defensible as a convenient means of gathering a large 25. Mechanism, —.57, —.87
sample of reasonably valid ratings, but it is well to 26. Operational definition, —.57, —.85
acknowledge certain inevitable kinds of bias in our data. 9. Biological determinants, —.43, —.37
Some theorists tend to be known primarily through second- 1. Learning, —.39, —.67
ary sources. To the extent that raters depend on these, 1.5. Immediate external determinants, —.35, —.59
they may be using a common fund of misinformation or
slanted interpretation. Permitting a free choice of theorists Positive theorists: McDougall, Jung, Brentano, Adler,
minimizes this source of bias but does not eliminate it. Piaget, Fechner, Janet, Hall.
The theoretical predilections of the rater constitute another Negative theorists: Estes, Watson, Pavlov, Spence,
source of bias, one whose effects are most subtle and Skinner, Miller, Hull, Guthrie, Harlow.
systematic when a pervasive contemporary outlook is
involved. Because of these and certain more trivial in-
At the positive pole, Factor 1 displays a pattern
fluences, some distortion is to be expected in mean values that might be described as subjectivistic, mental-
derived from the present data. Fortunately, the cor- istic, phenomenological, or psychological. The
relation coefficients used in this study and the values negative pole might be described as objectivistic,
derived from them are less likely to be affected system- physicalistic, positivistic, materialistic, or be-
atically by the sorts of bias we have just noted.
havioral. Perhaps the best label for the factor as
DIMENSIONS OF THEORY a whole would be Subjectivistic versus Objectivistic.
For each theorist the ratings from different cor- Factor 2
respondents were averaged, and a basic score matrix 14. Total organization, .96, .91
of order 54 X 34 was thus obtained. The inter- 28. Holism, .95, .94
variable correlations yielded by this matrix were 15, Uniqueness of individual, .65, .75
21. Naturalistic observation, .55, .68
subjected to centroid extraction, and six factors
were obtained. After an initial Varitnax rotation, 2
For score matrix and the most important matrices de-
the factors were rotated blindly to the best ap- rived from it, order NAPS Document NAPS-00004 from
ASIS National Auxiliary Publications Service, c/o CCM
proximation to simple structure. Both factor-
Information Sciences, Inc., 22 West 34th Street, New York,
pattern and factor-structure matrices—matrices of New York 10001; remitting $1.00 for microfiche or $3.00
loadings by and correlations with the factors, or for photocopies.
718 AMERICAN PSYCHOLOGIST

5. Observable behavior, .54, —.25 is consistently nomothetic, in the sense that it

10. Social determinants, .'12, .43 slresses the process or the slrticlure of behavior or
24. Einalism, .36, .80
21. Klcmcnlarisrn, .95, —.<>!
experience rather than (he behaving or experiencing
12. Influence of pas! experience, .05, --.26 individual or characteristics of the individual. The
23. Determinism, .51, - .66 negative pole emphasizes characteristics of the in-
25. Mechanism, —.46, - .76 dividual, but not necessarily in the sense of an
29. Nomothclic, --.28, —.36 idiographic approach. The essence of the former
Positive theorists: Goldstein, Kohler, Koffka, McDoagall, tendency is a certain mode of abstraction, but it is
Allport, Werthcimcr, Lewin, Rogers. not abstraction per sc, since the latter orientation
Negative theorists: Spcncc, Titchencr, Esles, Ebbinghaus, would admit the sort of abstraction found in per-
Hull, Wundl, Pavlov, Skinner.
sonality theory. For the present, this factor will
In Factor 2 we find a pattern that might he be labeled Transpcrsonal versus Personal.
called holistic, lotalistic, or molar, opposed by a
factor 4
pattern that might be called elementaristic, atom-
istic, or molecular. This factor may best be labeled 19. Statistical analysis, 1.05, .95
32. Quantitative description, .88, .97
Holistic versus Elementaristic.
31. Quantitative formulation, .79, ,93
The independent status of the first two factors is 30. Normative generalization, .76, .75
consistent with Brunswik's formulations. There is 18. Rigidly controlled experimentation, .48, .75
some vindication for Allport and other unidimen- 26. Operational definition, .47, .71
sionalists in the fact that there is a substantial 5. Observable behavior, .44, .55
11. Heredity, .41, .16
positive correlation between subjectivism and
29. Nomothetic, .26, .35
holism. On the basis of Brunswik's work, we 7. Emotion, —.48, —.54
might have expected the opposite, although it docs 6. Unconscious processes, —.29, —.60
not strictly follow from his analysis. It is to 20. "Armchair" speculation, —.28, - .73
Brunswik's credit that these two factors emerge 12. Influence of past experience, —.23, -.01
23. Determinism, —.23, .38
as the factors of greatest variance, but if is clear
17. Introspective reports, —.18, —.54
at the same time that we need several additional
factors to account adequately for the structure of Positive theorists: Estcs, Thurstonc, Spearman, Terman,
Spencc, Binet, Ebbinghaus, Miller.
psychological theory.
Negative theorists: Fraud, Janet, Goldstein, Charcot,
Jung, Wcrlheimer, Sullivan, Kohler, Koffka.
Factor 3
29. Nomothelic, .89, .69 This is the one factor that draws together the
34. Physical analogies, .87, .76 four variables that might be called quantitative.
13. Immediate external determinants, .69, .83 These are accompanied by variables suggesting an
2. Sensation and perception, .62, .43
emphasis on methodological precision and on con-
33. Hypothetical entities, .60, .01
18. Rigidly controlled experimentation, .50, .75 tent that lends itself to quantitative treatment. At
1. Learning, .31, .36 the negative pole we find a suggestion of procedures
16. Persisting traits, —.66, --.84 and content that do not readily permit quantitative
IS. Uniqueness of individual, —.57, --.76 treatment. This factor will be called Quantitative
30. Normative generalization, -'.40, —.19
versus Qualitative. The label, like the factor itself,
10. Social determinants, —.29, —.44
7. Emotion, —.26, --.50 is somewhat more sharply definable at the positive
6. Unconscious processes, - .21, - -.65 pole.
Positive theorists: Koffka, Kohler, Hull, Esles, Werlhei- Factor 5
mer, Skinner, Lashlcy, Titchencr, Miiller.
3. Motivation, .98, .94
Negative theorists: Rorschach, Binet, Adler, Jung,
12. Influence of past experience, .80, .67
Tcrman, Janet, Allport, Charcot, Hall.
33. Hypothetical entities, .73, .47
Factor 3 makes good sense psychologically, but 1. Learning, .57, .42
it is difficult to label satisfactorily. It could be 10. Social determinants, .55, .73
7. Emotion, .49, .67
called experimental versus clinical, but these terms 29. Nomothetic, .38, — J 4
overemphasize the methodological expression of the 6. Unconscious processes, .38, .45
factor. At the positive pole we see a pattern that 8. Self-concept, .39, .51
 DIMENSIONS ov PSYCHOLOGICAL THEOUY 719

24. Finalism, ..i2, .34 factors yield the following loadings on the first-
1C. Persisting trails, .22, .52 order factors: Factor A: .82, .75, .00, -.88, -.14,
2. Sensation and perception, -..(2, — .59
.14; Factor "B: - .10, .13, -.72, .00, .64, .36.
t7. Introspective reports, - .14, —.22
32. Quantitative description, .16, .3K Factor A displays a subjectivistic, holistic, and
30. Normative generalization, - -.IS, —.23 qualitative trend opposed to an objectivistic, ele-
Positive theorists: McDougall, Jung, Adler, Mowrer,
mentaristic, and quantitative trend. This dimen-
Sullivan, Freud, James. sion clearly coincides with Allport's Leibnitzean-
Negative theorists: Titchcncr, Wundt, Mach, Fechncr, Lockean dichotomy, and it brings together the basic
Werthcimer, Ebbinghaus, Spearman, Kiilpc. constellation which Ansbacher attributed to holism.
The variables associated with the positive end A possible general label would be Synthetic versus
of Factor 5 show a concern with ongoing processes Analytic.
or with things that tend to produce processes or Factor B contrasts a dynamic personal approach
change. The negative pole is less well defined, but with a static transpersonal approach, with an
suggests more emphasis on features that might be emphasis on internal or biological sources of be-
considered static or on methods that might be used havior tending to accompany the former. The
to isolate such features. A fairly accurate label factor is somewhat reminiscent of the clash between
would be Dynamic versus Static, though these terms James and the experimentalists of his day. Ft
have become a bit too value laden to be entirely might be broadly characterized as Functional
satisfactory. versus Structural.
If the labels suggested for the second-order
Factor 6 factors appear to lack precision, it must be remem-
9. Biological determinants, .80, .59 bered that we are dealing with influences whose
11. Heredity, .78, .75 effects tend to be dilute to the extent that they are
34. Physical analogies, .40, ,10 general. This rule holds as we move up one more
21. Naturalistic observation, .38, .56 stratum. It so happens that the two second-order
10. Social determinants, —.39, —.1-1
1. Learning, —.21, --.35
factors have a correlation of .55. To this extent,
8. Self-concept, -.16, .19 we may think of them as sharing a still more gen-
26. Operational definition, —.17, —.42 eral, but weak, dimension that might be designated
13. Immediate external determinants, —.16, --.40 fluid versus restrictive. The former pole sug-
Positive theorists: Gallon, Freud, Jung, Hall, McDougall, gests a basic predisposition to experience people
Cannon. and life in all their complexity in a rather re-
Negative theorists: Skinner, Tilchcner, Ebbinghaus, laxed fashion, while the latter suggests a tendency
Angell, Hull, Rogers, Watson. to deal with reality in a more controlling and
Here we find a biological emphasis at the posi- compartmental fashion, through restriction of at-
tive end, with such contrasting concerns as social tention and through isolation of entities and events.
influences and the learning process at the negative It has long been obvious that theoretical orientation
end. A constitutional emphasis apparently tends is at least partly a function of temperament. Per-
to go with the former and an environmental haps the broad factor we see here points to one im-
emphasis with the latter, but the nature-nurture portant personal source of intellectual outlook. Of
dichotomy evidently does not represent the central course, the personality variables associated with
focus of the factor. Rather, the essential dis- the first-order factors would offer a basis for more
tinction seems to be between the contrasting orienta- specific understanding and prediction. We must
tions—toward the internal sources of behavior and note that relatively few psychologists display either
toward the external sources—that find occasional extreme on the broad general dimension in a con-
expression in the nature-nurture controversy. A sistent way. In the present sample, James, Mc-
label that might capture the basic outlooks is Dougall, and Jung are fairly clear examples of a
Endogenist versus Exogenist. fluid orientation, while the restrictive pole is best
These six factors are intercorrelated, of course, represented by Skinner, Ebbinghaus, and Estes.
and an analysis of their correlations yields two Figure 1 provides a convenient overview of the
factors at a second stratum. When rotated to a entire system of variables and factors with which
position of simplest structure, the second-order we are dealing. They are shown in a double
720 AMERICAN PSYCHOLOGIST

Conscious processes)
Introspective report
Jnconscious process?
"Armchair" speculation|
Self-coiiceptJ
/o luntarisml

[Mechanism
[Ue termini sin
[Observable behavio
Total organisation!
N a t u r a l i s t i c observation!
[Operational definition
[Statistical analysTs
^ [Quantitative formulation
iuquencss ol individualj
Quantitative description
Normative generalization
[Social aeterminants|'
[Rigid Control
[Sensation 6 perception
[Immediate external determinants
Persisting traits]
Emotion
[Physical analogies
INomothctic approach

•lotivation|
Hypothetical entities]
Influence of past
Biological determinants]
leredityl

FIG. 1. A bipolar hierarchy of theoretical variables. (The variables shown in the middle
are relatively specific, while those on the left and right sides represent more general and mutually
opposing trends.)

hierarchy in which the converging lines represent variables. A clustering technique is preferable, and
the major links between a given set of trends and Ward's hierarchical grouping procedure (Ward,
those on a higher level of generality. The two .1963) was chosen for the purpose.
sides of the diagram represent opposite trends. In Since the score distributions varied somewhat
this scheme, of course, information is sacrificed from one variable to another, the first step was to
for the sake of simplicity. The crisscrossing of convert the values in each column of the score
lines is minimized because connecting links are matrix to a normalized standard form. All theorists
shown for only the highest combinations of loadings were then intercorrelatcd, and the correlation
and variable-factor correlations. Furthermore, the coefficients were transformed to d2 values. Ward's
interrelationships among the elements of a given procedure was then applied. This procedure es-
stratum cannot be properly depicted in a two- sentially performs successive groupings, according
dimensional diagram. This is particularly true to the similarities among elements and clusters
at the level of the original variables. In all, the formed from them, until all original elements are
hierarchy provides a convenient way of conceptual- grouped into a single cluster. From the record
izing certain prototypic theoretical perspectives. of the process, one can construct a hierarchical
For a variety of less common patterns of orienta- diagram that reflects the relative proximities in a
tion, it is somewhat less useful. test configuration (or, as in the present instance, a
person configuration).
TYPES OF THEORISTS The results are shown in Figure 2. The cluster-
Since all the theorists employed in this study are ing does not conform neatly to recognized "schools"
widely known, it seemed worthwhile to examine of psychology, but it is clearly meaningful. It is
their interrelationships more directly by some kind readily interpretable in terms of patterns in the
of Q-technique analysis. Factor analysis, however, factors discussed above. Thus, the cluster formed
is an inappropriate tool for this purpose. It yields by the first six theorists in the diagram is personal
more obscure results in Q technique than it does in and qualitative and moderately subjectivistic and
R technique, since persons are more complex than dynamic. The next five theorists display more
 DIMENSIONS OF PSYCHOLOGICAL THKOKY 721.

S. FREUD
P. JANET
If the factor scores for each set of 10 theorists
K. HORNBY
II. S. SULLIVAN
are averaged, we obtain the results shown in
J. M. CHARCOT.
II. RORSCHACH Table 1. The clearest trend by far is for Factor 1,
W. JAMES
If. McDOUGAl.L which shows a progressive increase in objectivism.
C. G. JUNG
K. GOLDSTEIN This agrees with Brunswik's hypotheses. An "ob-
F. BRENTANO
J. DEWEY jective" approach to methodology, of course, is
G. ». AI.LPOKT
A. AD LEU
C. R. ROGERS
better represented by Factor 4, for which the
J. PIAGET
G. S. tIAU, -
progression is basically U shaped. The marked
J. tt. ANGBLL
M. HERTHE1MER
upsurge in quantitative orientation in recent decades
W. K811LER
K. KOFPKA is consistent with Brunswik's views.
K. LEUIN
C. L. HULL From Brunswik's formulations, we might also
K, W. SPENCE
K. K. ESTES expect an upward trend in Factor 2. Instead, we
N. E, M I L L E R
I. P. PAVLOV find an inverted U-shaped progression, with a peak
J. B. WATSON
C. R, GUTIIRIE for holism in the 1920s. The remaining three
8. I'. SKINNER
E. t. T110RNDIKE
E. C. TOLMAN-
factors display more irregularity. Factor 3 shows
0. H. MOWKER —
K. S. LASIiLF.Y
an initial U trend settling in a plateau on the
D . 0. HEBB
C. S. SIIERRINGTON
transpersonal side. In Factor 5 there is a nearly
H. 1'. 1IAKLOW
\1. B. CANNON monotonic progression from the static to the dy-
W. WUNDT
E. B. T1TCIIENF.R namic. Initial fluctuation in Factor 6 culminates
0. KUI.PF,
II. von UELMHOLT7, in an endogenist peak between 1910 and 1919,
G. T. FECHNER
E. MACH followed by a monotonic trend toward the opposite
H. EBBINGHAUS
G. E, MfJLLER
E. BRUNSWIK
pole.
L. I. THURSTONE
C. E. SPEARMAN
If we examine successive decade averages in the
J. McK. CA1TELL
A. BINET
scores for specific variables, we find trends that
L. M. TERMAN
F. GALTON are largely parallel to those of the most heavily
R. S. WOODWORTH
loading factors. Both at the factor level and at the
FIG. 2. The hierarchical grouping of theorists. (The variable level, we find progressions that accord with
relative positions of points of juncture are a function of what is generally known regarding our salient
the degree of similarity of theorists and clusters of theorists
—the farther to the left the point is, the greater the theorists. At neither level is there a unified trend
similarity.) from which we can confidently extrapolate to future
developments. Perhaps our time range is too short
holism and a bit more subjectivism. The grand for such a purpose.
cluster formed by the first 22 theorists may be It is conceivable that unidirectional change is
broadly described as synthetic. This kind of truly characteristic of some factors, and we see
analysis could furnish valuable insights into pat- clearest evidence of it in Factors 1 and. S, but it is
terns of influence and interactions within various
theoretical traditions. TABLE 1
MEAN FACTOR SCORES TOR LEADING THEORISTS
HISTORICAL TRENDS IN THEORY WITHIN DECADES: 1880-1959
In our earlier study, theorists were rated by
decades in order that an examination of temporal
changes might later be made. For this purpose, we , 2 3 4 s 6
focused on the scores for the 10 theorists whose
contributions within each decade were deemed most 1880-1889 56.1 46.1 51.8 50.4 42.7 51.0
18904899 55.5 47.4 47.2 47.8 46.8 53.0
important. Such a group is obviously not fully 1900-1909 50.5 47.8 48.7 48.1 47.6 51 .5
representative of the great mass of conflicting ac- 1910-1919 49.6 53.0 48.3 46.2 52.4 55.1
tivities occurring in any one segment of time, but 1920-1929 46.2 54.8 55.4 45.4 49.9 S2.6
it tends to embody those tendencies which at- 1930-1939 43.9 51.9 55.1 48.7 53.4 48.5
1940-1949 42.1 49.4 56.3 56.2 53.8 43.7
tract greatest attention in each period and which
1950-1959 39.4 46.6 55.5 57.4 55.0 42.3
have the greatest influence on subsequent develop-
ments. e.— All values arc in T-scoiX' form.
722 AMERICAN PSYCHOLOGIST

more likely that we are simply looking at segments sider in any temporal analysis is that opposing
of long-term cycles. Surely cyclic change is a more movements do tend to operate concurrently, not
reasonable expectation for most dimensions, for merely in succession. For this reason, the mean
the more a given subject matter or perspective is value on a variable or factor for a given decade
neglected, the more it is likely to be needed as a may be somewhat misleading. A measure of vari-
corrective for the emphases that have displaced ance applied to a larger sample of representatives
it. An irreversible trend toward objectivism in for each period might better indicate the salience
psychology would surely lead to substantive sterility of a particular issue or dimension at any point in
within this science. The desirability of cyclic change time. There is much that remains to be illuminated
is clear in most of the factors. There are certain by the appropriate application of multivariate
things to be gained, for example, both from concern methods to problems in the history of psychological
with totalities and from attention to particulars. theory.
A temporary shift along the endogenist-exogenist REFERENCES
continuum may become fruitful by virtue of a fresh Ar.Li'OKT, G. W. Becoming: Basic considerations for a
innovation in physiological thought or a new method psychology of personality. New Haven: Yale Uni-
of studying organism-environment interaction. versity Press, 1955.
There are some forms of change for which our ANSBACI-IER, H. L. On the origin of holism. Journal of
Individual Psychology, 1961, 17, 142-148.
present methods are not suitable. There is little
HRUNSWIK, E. The conceptual framework of psychology.
doubt that some historical developments can be In, International encyclopedia of unified science. Vol.
viewed in terms of a succession of triads, wherein 1, No. 10. Chicago: University of Chicago Press, 1952.
the reconciliation of a given polarity is followed COAW, R. W., & ZACONA, S. V. Contemporary ratings of
by a new division of viewpoints. On this basis, we psychological theorists. Psychological Record, 1962, 12,
315-322.
might expect a succession or a concurrent inter-
JAMKS, W. Pragmatism. London: Longmans, Green, 1907.
action of the two polar constellations of a factor MURRAY, It. A. Hxplorations in personality. New York:
to be followed by a drastic realignment of vari- Oxford University Press, 1938.
ables, so that new patterns of covariation emerge ROGERS, C. R. Two divergent trends. In R. May (Kd.),
and a new factor replaces the old one. This is Existential psychology. New York: Random House,
1961.
only one of several ways in which we might envision
WARD, J. H. Hierarchical grouping to optimize an ob-
the transmutation of factors over time. jective function. Journal of the American Statistical
Perhaps a more important complication to con- Association, 1963, 58, 236-244.