University of Virginia Library

IV. Part IV

DISCUSSION

By Clyde Kluckhohn

The Position of Bc 51.—It will have been noticed that the Pecos
classification has been used in the distributional analyses but not in
the description of Bc 51. In other words, it has been used only when
others had assumed the responsibility for the assignment of a site or
a part of a site to a particular category in the Pecos classification. The
reluctance to assign a label to Bc 51 is not to be understood as based
upon a conviction that the Pecos classification has not been useful in
the development of Southwestern archaeology. But close study of the
recent literature of the subject seemed to show some confusion on certain
issues of classification and suggested that an unqualified assignment
of this site might lead some later comparative student to make
equations not altogether justified by the facts. The concrete data from
Bc 51 bring out particular classificatory difficulties sufficiently sharply
that a fairly extended discussion will, perhaps, be profitable.

At first glance the matter may appear to be simple enough. As
is well known, Roberts has, on several occasions within the last few
years, published lists of traits diagnostic of the various Anasazi subdivisions,
and these lists appear to have been widely accepted. One
might think, then, it was merely a question of determining whether the
material from a given site conformed to a particular set of specifications.
But, if one is working with a specific assemblage of data and
endeavoring to follow out operations which are fairly precisely defined,
a number of questions arise which would seem to require clarification:
is it absolute presence or absence of the criteria which count
or merely predominance—or does the answer to this question vary in
the case of various traits? Must the culture or "culture period" check
with all or with how large a majority of the diagnostics? Are certain
of the criteria indispensable and others not? That is, suppose 8
criteria are taken as diagnostic of Pueblo I in the Chaco and suppose
it is agreed that at least 6 of these must check (so far as there is evidence
available on the trait in question), will we still call a site Pueblo
I if 6 of the 8 criteria are found indubitably associated with a masonry
type or a pottery complex which has been accepted as diagnostic of
Pueblo II? Cases of this general sort are not unknown to experience.

Observation of the actual operations of archaeologists suggests
that in many cases the classification of a site is actually made on the
basis of pottery complex or architectural style (including masonry
type) alone. If this fact is explicitly stated, this procedure may well
be the most convenient and quite unobjectionable. If, however, there

is assertion or implication that the classification has been made on the
basis of total culture complex, some confusion results. For what occurs
is that other culture elements found associated with the critical pottery
complex or architectural style are simply dragged in after the
crucial step has been taken. If we are really operating with pottery or
masonry-architectural complexes (or a combination of these two)
only, it would be in the interests of clear thinking if this circumstance
were brought into the open, either through terminology or explicit
statement. It seems possible that classificatory operations frankly
based solely upon these apparently somewhat more sensitive and more
consistent criteria would be the most useful. The associated culture
elements (not used in cultural classification) could then be studied
apart from the prejudice of a question-begging nomenclature and,
after the trends toward uniformity had been unequivocally ascertained,
the operations for definition which was truly in terms of total
culture complex could be rigorously set up.

It would, then, be necessary to make clear how great a proportion
of traits otherwise regarded as diagnostic of, let us say, Pueblo II,
could be admitted seriatim in a Pueblo I site. It would, likewise, be
imperative to state if any criteria are to be weighted as of greater
importance and which differentiae are to be applied first. Because, as
in physical anthropology, two investigators can use the same diagnostic
traits in making a classification and yet get different results,
depending on the order in which they are applied (with resultant eliminations).
It follows, also, that the relative significance to be attached
to positive and negative evidence would have to be specified.

Finally, there is the vexatious "time problem." Are tree ring
dates to have a part in our assignment and, if so, how important a
place? Roberts has written: "It should be emphasized that these
designations apply to the complex and not to a single element or series
of years."[1] Roberts appears to adhere quite consistently to this
position. For example, he has recently insisted that a particular Chaco
masonry type (Slab Base Rubble) is both Basket Maker III and
Pueblo I (i.e., he rejects masonry type as a binding criterion) for
"Judd's report on this house and the published pictures of the pottery
and other objects found in the structure clearly indicate that it belongs
to the Pueblo I period."[2] Even here, though, the use of the word
"period" suggests that the time factor, as well as culture complex,
enters in. And Roberts has elsewhere[3] indicated that he
realizes that terminology of this sort tends to acquire "a time connotation
as well as a descriptive meaning." Southwestern archaeologists,

undoubtedly, continually write and speak of "Pueblo II times" or
"the Pueblo I period." As a matter of fact, while most Southwestern
archaeologists seem to give verbal assent to Roberts' proposition that
the Pecos classification is merely a descriptive categorization of assemblages
of traits, evidence is not lacking that dating slips in as a covert
criterion. Indeed, within the last few years, a few archaeologists,
apparently disturbed by the lack of objectivity of other procedures,
avowedly classify sites primarily by their tree ring dates. For example,
we find Baldwin writing ". . . it was occupied from 1150 to 1320
A.D., thus belonging to Pueblo III . . . and the early part of Pueblo
IV."[4]

In short, some archaeologists, at present, use the labels of the
Pecos classification in accord with pottery or architectural complexes
found present, others endeavor to take account of total culture complexes,
still others make the distinctions primarily on the basis of
tree ring and documentary dates—with multitudinous ill-defined combinations
of these three alternatives. The real difficulty is that the
criteria are not consistently used. Either they are mingled (in a
manner which is not made explicit) or one set is used on one occasion
and a different set on another. Hence, various questions are begged by
the assignment of such labels. A worker gets no (or an inadequate
number of) tree ring dates at a given site but labels it "Pueblo I."
Experience shows that comparative students are all too likely to
assume that the culture of this site can be chronologically equated
with that of dated sites, perhaps in quite different regions. The converse
error is, perhaps, less frequent but also occurs. It is probably
true that the more competent professional Southwestern archaeologists
are fully aware that "Pueblo II" has dates which range over a wide
spread in various areas, but archaeologists specializing in other regions,
and ethnologists, are more likely to fall into the fallacy of "one
culture complex, one period." Thus, Dr. Parsons, in her recently
published monumental work on Pueblo religion, writes without
qualification "In the archaeological period called Pueblo II and dated
about 875 to 1000 . . . "[5] The fact of the matter is that one may find
sites assigned to Pueblo II dated in the Tree Ring Bulletin from as
early as 817 to as late as 1144.[6] The facts of Southwestern archaeology
should surely not be presented in a terminological form resulting in
confusions of this sort on the part of experts in peripheral fields. Unfortunately,
Roberts, in one of his surveys for the non-specialist, does
not help matters when he states that Pueblo II dates approximately
from ". . . 875 to 950, longer in the peripheral districts."[7] It is true

that elsewhere in his article he says, "While the progression of stages
infers a certain degree of contemporaneity between sites of the same
horizon, it does not, necessarily, mean that they will fall within identical
chronological dates."[8] Nevertheless, it seems certain that the
reader, whose primary professional interest is not Southwestern
archaeology, would gain an impression that the spread of dates for
particular culture complexes is considerably less than is in fact the
case.

Until one is prepared to say relatively unequivocally: "When I
say Pueblo II, I mean a culture distinguished by the presence of the
following traits . . ., by the absence of the following . . ., by the
relative preponderance of such and such traits as opposed to such and
such others in proportions which approach statistical constancy" or
"When I say Pueblo II in a given area, I mean sites or distinguishable
portions of sites which give no dates earlier than . . . and no dates
later than . . . ," confusion is quite likely to result unless any assignment
of the labels of the Pecos classification is guarded and unless
the basis or bases for such assignment is made fully overt. In the case
of sites where tree ring dates or widely varying masonry or pottery
styles indicate the likelihood that more than one culture complex is
represented, the need for assurance that the various features of the
inventory were truly associated in the usage of a particular people
becomes particularly urgent. Association in the refuse mound or room
fill of such a site may well be at least once removed from the actual
historical complex. Even in cases where articles are found indubitably
grouped with a single burial or on the floor of a room, the
possibilities of error from heirloom pieces or intrusion have, perhaps,
been underestimated and, on close examination, the quantitative basis
for some generalization appears hopelessly inadequate.

So much for a consideration of these problems in the abstract.
Let us now look at the Bc 51 facts in the light of them. An archaeologist
who casually inspected the site and the material from it would,
if he ventured an opinion, be likely to describe it as a Pueblo II site.
If one were able to analyze the intellectual operations performed in
making such an assignment, they would also certainly reduce themselves
to two or perhaps three. First of all, since Bc 51 is not, on the
one hand, "a great terraced communal house of many rooms," nor, on
the other hand, of slab construction, there would be an unwillingness
to assign it either to Pueblo III or Basket Maker III. In the second
place, the pottery types (Exuberant Corrugated and Escavada Black
on White) which are generally the most numerous types in rooms and
kivas, are types which competent pottery specialists have considered
to be diagnostic of Pueblo II. A third general type of consideration
might also enter into the judgment. Certain types of complexity and

elaboration of material culture, which one associates with Chaco
Pueblo III sites like Pueblo Bonito, were not found here. There are,
however, certain traits present which would seem fairly definitely to
rule out Basket Maker III and, possibly, Pueblo I.

Probably this is not an unfair dissection of the rough and ready
process of reasoning which many archaeologists would go through in
making a first approximation to an assignment of the site in terms of
the Pecos classification. But now let us examine the matter with somewhat
greater attention to detail. Suppose, to start with, we use Roberts'
widely accepted "Survey of Southwestern Archaeology" in trying
to decide whether this is a Pueblo I or a Pueblo II site. In five of
his sets of diagnostic traits (Sandals, Basketry, Textiles, Weapons,
"Other Traits") no features are listed which set off Pueblo I from
Pueblo II. This leaves: Crania, Pottery, Houses. Roberts suggests
that round crania are proportionately more numerous in Pueblo I
sites. However, recent and as yet unpublished research by Dr. George
Woodbury and Dr. Carl Seltzer indicates that no clear-cut distinction
can be made between the crania of sites which have been assigned to
Pueblo I and to Pueblo II. And so we are left with Pottery and Houses
as diagnostic criteria.

So far as published evidence goes, the conviction that Escavada
preceded such types as Gallup and Chaco Black on White in predominant
popularity rests on the data from a single site,[9] except that the
latter types gain in prominence in the upper levels at Bc 50, Bc 51,
and Łeyit Kin. We can probably grant that the available data indicate
that Escavada is a type of somewhat earlier popularity. Dr. Hawley's
Escavada Pottery Complex seems quite certainly to be later than the
Red Mesa Pottery Complex and earlier than other pottery complexes
found in the Chaco. But is there satisfactory evidence that the Escavada
Pottery Complex may be regarded as determinant for a total
culture complex? Reserve on this point surely seems proper. It would
be hard to make an unimpeachable case for clear differentiation of the
inventory of artifacts associated with the Red Mesa Pottery Complex.
And the prominence of Exuberant and Escavada must not obscure the
fact that types the floruit of which is given a Pueblo III provenience
by Dr. Hawley and trade types assigned in their indigenous areas to
dates usually considered to be definitely Pueblo III, are present in
appreciable numbers in more than one locus and level. Also, we must
be vigilant against an infinite regression in reasoning here. If, because
a Pottery Complex appears to precede one or more later Pottery
Complexes associated with the great communal houses typical of
Pueblo III in the Chaco, we take the first Pottery Complex as the
determinant of Pueblo II in the Chaco, it is mere tautology to cite

these pottery types as additional evidence for a site being Pueblo II.
It has been so defined.

Turning to "Houses," Bc 51 does not altogether conform to Robert's
list if one takes it literalistically. He says, "These dwellings . . .
contained from six to fourteen rooms." Bc 51 has more than 19 rooms.
Roberts writes: "Usually at the south or southeast side, detached from
the building, was a subterranean ceremonial chamber." One kiva at
Bc 51 was at the southwest side, another at the northeast, but, after
all, Roberts has qualified here by "usually." Kivas 3, 4, and 5 are not
detached from the building, although kiva 5 may have been when built.
Land contours may, of course, account for certain variations. There
are 6 kivas at Bc 51. It is more than plausible that not all of these
were used simultaneously but it hardly seems likely that only one was
used during a single time interval. This is not written with a view to
carping at Roberts' excellent summary, but it seems worthwhile to
point out the difficulties when one attempts to apply the minutiae of
definitions with precision. Residually, at least, Bc 51 is Pueblo II in
house type.

Small difficulties and reservations aside, this much seems certain
and important: if we do label Bc 51 as a Pueblo II site, we are so doing
on the basis of two sets of criteria (Pottery and Houses) alone. It is
not an assignment on the basis of total cultural complex except, perhaps
in a grossly negative way (that is, we did not find atlatls nor
objects of European manufacture, for example.)

To be sure, some of the problems which have been raised disappear
if we adopt the newer terminology suggested by Roberts and
say simply that this is a Developmental Pueblo site. A number of
workers have pointed out that Pueblo I and Pueblo II were, perhaps,
the most dubious categories in the Pecos classification and the very
fact that so few sites have been assigned by their excavators to Pueblo
II suggests that this complex may have been an ill-defined or needless
category.[10] At best, it has been treated in practice as something between
a Platonic Idea and a residual category.

Even, however, if we call Bc 51 Developmental Pueblo, difficulties
connected with the time issue remain. The only tree ring dates obtained
from this site fall relatively late within the time interval during
which the large structures across the canyon, almost invariably
assigned to Pueblo III, seem to have been occupied. It is only fair
to recall that these dates came from a room which had been partially
refaced with masonry of a type assigned by Dr. Hawley to Pueblo III.
It is possible, indeed, that they represent merely the reoccupation of a
single room in Bc 51 by persons from across the canyon. On the other

hand, even the date of 922+, obtained from Bc 50, is as late as that of
cutting dates of logs from Pueblo Bonito, Chetro Ketl, and other
"Pueblo III" Chaco sites. The possibility may be granted that logs of
earlier dates from the "Great Pueblo" sites might represent reused
logs obtained from abandoned sites of different architectural type. In
any case, however, the principal period of occupation of most Bc 51
superstructure rooms practically certainly falls materially later than
922. In Dr. Hawley's opinion "sometime between 975 and 1045" would
be a fair estimate as the building period for Bc 51. Unless various
supposedly trade pottery types (such as Sunset Red and McElmo)
appear in the Chaco considerably earlier than in their putatively indigenous
areas, Bc 51 must have been extensively occupied at least as late
as about 1000.

In short, a minimum statement would be that human beings were
almost certainly living in Chetro Ketl and Pueblo Bonito during at
least part of the time that Bc 51 was occupied. Now we do not as yet
have a complete report on any of the Great Pueblo Chaco ruins and
it would be a mistake to assume too readily that the cultural remains
from, say, Pueblo Bonito, are altogether homogeneous. Nevertheless,
Pueblo Bonito and other Chaco sites of similar architectural type[11]
are commonly taken (without qualification) as typical Pueblo III or
Great Pueblo. In other words, during a certain time interval, the cultures
on the north side of Chaco Canyon were Great Pueblo, the at
least partially, contemporary ones about a mile away on the south side
were Developmental Pueblo.

If this be so, either our definitions fail somehow to correspond to
the historical actualities, or there is some rather special explanation.
One possibility of the latter sort suggests itself. In looking at all of
the facts from Bc 50-51, one is presented with two general alternatives
of interpretation. There may, on the one hand, have been essentially
continuous occupation from the time of the pithouse dwellers
forward. The development may have been unbroken and largely autochthonous,
and those who last lived in the northern rooms of Bc 51
may have been the lineal descendants, culturally and physically, of
the carriers of the Lino Gray pottery complex at Bc 50-51. Some of
the archaeologists with whom the writer has discussed the evidence
prefer this as the most economical hypothesis. They would also favor
the view that there was continuity between the cultures which seem
to have existed side by side, for a time, on the north and south sides of
the canyon. Apart from architecture-masonry and, perhaps, pottery
types, the cultural differences between Chaco Great Pueblo and Developmental
Pueblo sites appear, so far as present knowledge takes us, to

reduce themselves largely to the greater richness in cultural inventory
of the former and to the presence of certain articles (such as the
well-known Bonito mosaics) implying a culture of greater complexity
where there were more resources and more specialization of labor. On
this view, the inhabitants of Bc 51 were either "poor relations" or
conservatives who refused to adopt the progressive architectural styles
of their congeners across the canyon.

This may well be the correct interpretation. But, while there is
no proof that Bc 50-51 was not continuously inhabited, there is equally
certainly no proof that it was. And it is well known that abandoned
habitation sites in the Southwest have often been later reoccupied by
quite distinct peoples. Let us therefore develop, rather boldly, an
alternative hypothesis. Speculation in science is dangerous when it
is not clearly designated as such. But frank speculation is sometimes
valuable in preventing the closure of the mind to the range of equally
possible interpretations of a given set of facts. And if alternative
explanations are admitted at a given stage in research, the next stage
of investigation will (or should) be planned broadly enough to test
the choice. Whereas, if the only possibility which is envisioned is that
which first occurs, the range of investigation may be so narrowed that
the advance of knowledge is retarded.

Let me start with a statement which can necessarily hardly rise
above the level of personal opinion. In terms of what we know about
Pueblo cultures, both archaeologically and ethnologically, it does not
seem to me altogether plausible, at the moment, that the people who
lived in the sites on the south side of the canyon were carriers of
precisely the same cultural tradition as the inhabitants of the great
communal houses. The very circumstance that the "small house sites"
seem to be almost entirely on the south side of the canyon would itself
appear to be significant. At least, the contemporaneity of the two architectural
styles militates against either a purely developmental or a
purely physiographic explanation of the localization of the two
architectural types. It is possible that at farming colonies one would
find certain of the more valuable articles unrepresented, but are these
sites far enough away to be farming colonies of the sort known among
the modern Pueblo? Moreover, the fact of the numerous ceremonial
chambers and the fact that virtually all of the non-perishable artifacts
necessary for existence have been found, would seem to make
possible the inference that the dwellers led an essentially independent
life. I am aware, of course, that the material culture of the modern
Hopi towns of First and Third Mesas, for example, is not identical in
all important particulars. But I know of no case where different contemporaneous
villages of the same "tribe" had differing major architectural
and masonry styles or such marked disparity in richness of
total cultural inventory.

Another guess than that the inhabitants of Bc 51 were the "poor
relations" of the people who were the carriers of the "Pueblo III," culture,
might be that they were (perhaps in part) migrants from another
region, representatives of a related but somewhat less advanced cultural
heritage drawn to the Chaco by the prosperity of its inhabitants,
or, conceivably, by the reputation for magnificence and power of their
ceremonialism, or by the protection which these populous towns could
offer (or by a combination of any or all of these factors). If we may
judge at all from documented Pueblo history, such a movement in a time
of stress or trouble is altogether in accord with the configuration of
Pueblo behavior patterns. Roberts[12] and others have suggested migrations
to the Chaco from southwestern Colorado from fairly early times
on. Dutton has suggested this interpretation for Unit III at Łeyit
Kin.[13] Such a possibility for Bc 51, perhaps, gains slightly in credibility
from the fact that McElmo and other wares, supposedly indigenous
to that region, appear to be appreciably more prominent at Bc 51
than at Chetro Ketl, the only Chaco Great Pueblo site from which the
pottery has been reported in detail. This is, clearly, sheer speculation,
but it is speculation which will permit of some testing in the light of
future evidence. Surely, the apparently synchronous presence within
less than a mile of one another of sites which would be assigned to
Great Pueblo and Developmental Pueblo requires explanation. The
almost contiguous presence of migrants from some distance presents
no circumstance contrary to what we know of Pueblo history. In fact,
the Laguna group at Isleta, the Tewa on First Mesa, the plausibility
of Keresan-speakers in the Jeddito Valley make it altogether conceivable
that the inhabitants of Bc 51 might even have spoken a different
language from the dwellers in, say, Pueblo Bonito.

Probably the single fact of greatest general import which has
emerged thus far from the Bc 50-51 excavations is that the various
stages recognized by the Pecos classification (and very commonly referred
to as "periods") do not, necessarily, represent separate and
clear-cut time periods, even in the same geographical locality. Here it
can hardly be merely a question of a brief overlap. Tree ring dates
and pottery types both make it almost certain that cultures which most
archaeologists would designate as Developmental and Great Pueblo
existed for a considerable time within a very short distance of one
another.

In view of this fact (and of other difficulties which have been
mentioned) one is inclined to wonder whether the Pecos classification
has not, after all, outlived its major usefulness. That it has been
most useful in the ordering and systematization of the multifarious
data of Southwestern archaeology, no sensible person would, I think,

question. But we must remember, as Whitehead has so often reminded
us, that a classification is, at best, "a half-way house." A classification
is useful so long as the facts fall without violence into it. So soon, however,
as their greater bulk, greater complexity, or greater subtleties
of discrimination make the classification a Procrustean bed into which
the maimed and helpless facts are forced, the classification should be
abandoned or radically modified. As Kidder has observed: ". . . our
investigation has now reached a point at which formal classifications,
such as the Pecos nomenclature are not only of lessening value, but are
often, as in the present case, positively misleading.[14] At very least the
Pecos classification should take explicit account of the differing periods
of development in different areas and of varying genetic sequences in
various regions of Anasazi culture. To be sure, the concrete difficulties
which have been pointed out here arise out of the facts from but a
single locality. The possibility that generalizations of their implications
for Southwestern archaeology as a whole have been exaggerated
is readily granted.

Possibly some may feel that a taxonomic system modeled on the
well-known McKern terminology for Midwestern archaeology might
best serve the double end of ordering the data and still keeping us
close to the observed facts. In the light of the demonstrable confusions
resulting from the time problem, this plan, whereby the data would be
analyzed descriptively with tree ring dates simply stated as additional
facts sounds attractive. On the other hand, some workers who have
used the McKern system report that the lack of a time dimension is
acutely felt. Therefore, the revision of the Gladwin classification
(really a fusion of Gladwin's system with that of McKern) which Dr.
Colton will shortly publish[15] (and which he kindly allowed me to read
in advance of publication) would appear a step in the right direction.

This classification would have the great advantage (for the situation
under discussion) of not exaggerating the distinctions between
Bc 50-51 and Łeyit Kin on the south side of Chaco Canyon and Pueblo
Bonito and Chetro Ketl, almost directly opposite on the north and
almost certainly lived in during the same period that considerable portions
(at very least) of the smaller sites were occupied. Under the
Pecos classification, one is compelled to assign the two sets of cultures
to utterly different categories. This seems a distortion of the
facts. Under the system which Colton will set forth, they would, presumably,
be classified merely as different foci of the Chaco branch.
This distinction would be analogous to that between modern Pueblo
"tribes." Since, for all of the impressive differences in architecturemasonry
and in some other features, the two sets of inventories show

such an overwhelming number of artifacts and articles of subsistence
in common that this terminology would seem to adhere more closely to
the contrast given by the data. This is congruent with one of the
generalities which appeared from the distributional analyses: cultural
similarities and continuities in the same geographical area are
most impressive. In addition to some striking differences, the number
of close parallels between the artifacts of Shabik'eshchee and Bc 51 is
rather amazing when one comes to total them up. The Colton system
will also, doubtless, make it easier for us to keep clearly in mind the
fact that single sites do not necessarily represent a single time interval
nor a single homogeneous culture. Our minds are all too prone to
utilize tags which relegate inconvenient complexities into the background.
It is so easy to say "Pueblo Bonito is a Pueblo III site" and
be done with it. The "component" and "focus" terminology seems
likely to bring out more clearly the probably somewhat intricate cultural
history of such sites than does the unilinear and limited set of
categories of the Pecos classification.

That Bc 50 (superstructure at least), Bc 51, and Łeyit Kin (Units
I and II at least) would be considered simply as components of a single
focus, appears probable. It seems likely that some rooms at Bc 50 were
built before most of Bc 51 and there is no proof that any rooms in the
two mounds were ever lived in during precisely the same years. But
of cultural differences which appear to be consistent and significant
there are few. Almost all distinctions are of this sort: no twilled ring
baskets were found in Bc 51 and no coiled baskets in Bc 50; no awls
of class 1a were found in Bc 51, although several were found in Bc 50;
one object of antler was found in Bc 51, none in Bc 50; no notched percussion
tools were found in Bc 51; the palettes found in Bc 50 were
larger. The proportion of extended burials is conceivably significantly
higher in Bc 51.

Now, clearly, such variations depend so much on negative evidence
and upon accidents of the sampling process that no case for cultural
differentiation can be built upon them. Perhaps the evidence affords
some slight indication of variation in cultural fashions which may well
be correlated with the hypothesis that the period of floruit of Bc 51
was somewhat later than that of Bc 50. But an archaeologist
would be hard put, indeed, to distinguish random samples of the
objects found at the two mounds. Almost the only contrast which
seems perfectly clear-cut and indicative of a distinct cultural pattern
is the consistent presence of turkey bones behind the fire screens in all
of the kivas of Bc 50! The range of variation, with respect to Łeyit
Kin, appears slightly greater, perhaps because of its possibly more
heterogeneous character. But the same general conclusion applies.

Throughout this section emphasis has been laid—probably overmuch—on
terminology and classification. A classificatory nomenclature

is indispensable if we are not to become lost in a welter of isolated
facts. This consideration must not, however, be allowed to obscure the
all important fact that we are dealing with the products of the activity
of human beings. Human activity notably fails to exhibit exceptionless
uniformities. In concluding this discussion of the position of Bc 51
stress should be laid upon the range of diversity exhibited within this
limited collection of material. Consider the kiva plans (Fig. 5). How
far, in respect of individual features, do they rigidly conform to a
single pattern? Even the southward orientation of the ventilator
shafts (which Dr. Kidder[16] has considered "a most stringent ceremonial
requirement" within the San Juan area) is far from constant. Surely,
all classifications can but, at best, express modal tendencies and must
be used purely heuristically, with constant awareness that they are
most crude categorizations of the human acts we are trying to
reconstruct.