3. Geology, Biology, and Linguistics. The metaphor
“linguistic paleontology” exploits a certain obvious
analogy between “dead” languages and fossils. (It
should be borne in mind that paleontology, the study
of fossils, belongs both to geology and to biology; this
fact leads us, when considering influences of geology
upon linguistics, to consider at the same time certain
influences of biology upon linguistics.) For one thing,
linguistics was concerned with normative questions to
a far higher degree than was biology. There was, for
example, a counterpart in linguistics but not in biology
to the literary “quarrel of the ancients and the mod-
erns.” This dispute—ably described in Chapter 14 of
Highet (1949) and in Wimsatt and Brooks (1966), pp.
214, 262, 437n., 523-24—concerned the relative value
of ancient languages and literatures as compared with
modern languages and literatures. Whereas there was
never a time in biology when the study of fossils was
more highly esteemed than the study of living plants
and animals, it was only after centuries of debate that
the study of living languages and literatures (written
or oral) came to be considered not inferior to the study
of Latin and Greek. And the debate was, in effect,
ended sooner for literature than for language: the
“progressive” view prevailed, very broadly speaking,
for literature already in the Enlightenment, but for
language not until romanticism. The chronicle is well
told in the various writings of the ardently progres-
sivist, and yet remarkably objective linguist Otto Jes-
persen (e.g., 1922).
The maxim, therefore, that we ought to interpret
dead languages in the light of the living ones was more
controversial in linguistics than its counterpart was in
biology. It was countervailed by the argument that,
as the living languages have degenerated from the dead
classical languages, to use the former as a guide is to
interpret the perfected and optimal form in the “light”
of the corrupted and obscured form. In biology a like
argument was heard only in a very few special cases,
for instance, in the opinion about ancient man that
“There were giants in those days.” The superior
importance of the classical languages was bolstered by
a judgment about the superior importance of written
language over spoken language. According to this
judgment, if we have only written records of ancient
languages, this is, to be sure, a loss, but not an essential
loss. In biology, per contra, it was generally recognized
that if, e.g., one classified fossil molluscs exclusively
according to properties of their shells, this basis of
classification, used for lack of anything better, was
forced upon us by the circumstance that only their hard
shells, and not their soft inner vital parts, got preserved.
The view attained in the nineteenth century, that we
lacked information about such “vital parts” of the
classical languages as their system of intonation, the
details of their pronunciation, and the full extent of
differences of dialect, social class, and style within
them, and that in drawing inferences about them we
should take as our models the living languages that
we could more fully observe—this view was in effect
uniformitarian, though there was no demonstrable
influence from geology.
Having seen that uniformitarianism involved taking
a position on the interpretation of those episodes of
the Bible that concerned geology, we may inquire what
positions linguistics took on the corresponding
linguistically relevant episode. Linguistics had no
Lyell, and the problem arises, Why not?
Several ingredients of an explanation present them-
selves. One of them—the normative aspect of linguis-
tics—has already been discussed. In the second place,
the hypothesis of uniform change had a very different
aspect in linguistics than in geology, and this for
several reasons. (1) Rate of linguistic change was less
easily quantified than, for example, rate of sedimen-
tation, glacial advance, or lava flow. The point is
not that it was less easily measured, but that the very
definition of what to measure was less easy. (2) If one
estimated the rate of language change impres-
sionistically, one might well have had the impression
that it was highly variable. In particular, it was
plausible to suppose an acceleration in times of social
upheaval, or of decline in level of education, etc. (3)
The mathematics of probability and statistics, which
could be put to good use in quantifying rate of linguis-
tic change, had not been sufficiently developed until
the early nineteenth century, and, when it was devel-
oped, it was not regarded by linguists as one of their
available tools.
In the third place, there are significant differences
between the linguistic aspects of the Confusion of
Tongues at Babel and the geological aspects of the
Creation and the Flood. The Babel story, taken
literally, says that by extraordinary, supernatural inter-
vention God replaced a linguistic unity by a linguistic
diversity that was sufficient to frustrate universal com-
munication, but this left room for great latitude of
opinion about the nature of the replacement. One
extreme possibility was that God maximized the
diversity; the consequence of this for scientific linguis-
tics would be that we could not by any set of regulari-
ties account for the relation between the lingua
adamica (whether this was Hebrew or some other
language) and all the now existing languages. The other
extreme possibility was that God minimized the
diversity; the post-Babelian languages, each of them
describable as gotten by a set of regularities from the
lingua adamica, differed from one another just barely
enough to frustrate universal communication, yet not
so much as to frustrate scientific understanding of their
relationships. This seems to have been Leibniz' view,
in keeping with his principles that God works every-
where by rules and that he achieves maximal results
with the minimum means. Thus both the so-called
“polygenist” view and the “monogenist” view can
claim compatibility with the Babel story. But neither
in Leibniz' time nor in Müller's nor in ours has any
great success been achieved in working out the
monogenist view in any detail.
As for the hypothesis that the pre-Babelian language
was Hebrew, linguistic science was not in a position
to refute this hypothesis definitely until about the
1860's, i.e., until the Comparative Method reached the
stage to which Schleicher brought it.
In the fourth place, linguistics was subject to more
constraints than geology. Even if relieved of heter-
onomy from Revelation, linguistics was still subject to
constraints external to it, from physics, from geology,
and from biology. Archbishop Ussher's chronology,
fixing the Babel episode at 2347 B.C., gave the linguist
only several thousand years to work within, as noted
by Sapir in his discussion of Herder ([1907], p. 117;
cf. Jespersen [1922], p. 28), but the natural sciences
did not allow him much more. Given the obvious fact
that man is of relatively recent origin, and given the
opinions prevailing in the nineteenth century about the
age of the earth and the durations of the several
geological periods, there were still only a few thousand
years available to the linguist for the origin and
diversification of language. Geology and evolutionary
biology were themselves oppressively constrained in
their available time span by current physics (Eiseley
[1958], Ch. 9); for instance, in 1893 Lord Kelvin
accepted an estimate of the earth's total age as 24
million years. Contrast this with the fact that by 1907
B. B. Boltwood, using the half-life principle of radioac-
tive decay, had arrived at an estimate 100 times that
of the earlier one (personal communication from Matt
Walton, 1963; not in Eiseley). Present-day estimates
agree more or less with Boltwood's value rather than
Lord Kelvin's, and the antiquity of Homo sapiens has
likewise greatly expanded.
In the fifth place, linguistics was preoccupied with
another task. Whitney (1867, p. 3), sketching the his-
tory of linguistics, spoke for the prevailing view when
he called the recognition of an Indo-European family
“the turning point in this history, the true beginning
of linguistic science.” The challenge of following up
this recognition led to devising the Comparative
Method (Bloomfield [1933], Ch. 18), and to refinements
of methods in historical linguistics. The grand project
of working out, by the historical and the comparative
methods, the detailed history of the Indo-European
family and also of various other families (Semitic,
Dravidian, etc.) occupied nearly all the energy put into
linguistics in the nineteenth century. But the historical
and comparative methods did not—except perhaps for
their indications that Hebrew was not the same as
Proto-Semitic—yield any result that challenged re-
ceived interpretations of the Bible. Besides the reason
already given—great uncertainty about the rate of
linguistic change—there was another, that needs to be
set forth rather fully. This other reason was that the
comparative method cannot settle the question
whether all human languages are descended, by
uniformitarianly acceptable processes, from a single
ancestor. The diversity of known languages is such, and
the imperfections of the method are such, that the
method breaks down before it reaches the end of our
quest, which is the beginning in time of language. This
limitation became clear to linguists in the second half
of the nineteenth century.
The reason for the incapacity is that the comparative
method does not admit of unlimitedly recursive appli-
cation. Taking historical, documented languages as our
input, we get as output Proto-Indo-European. And
taking other sets of documented languages as inputs,
we get Proto-Semitic, Proto-Finno-Ugric, Proto-
Algonquian, and so on. The idea occurs to us that we
could treat these reconstructed proto-languages as
inputs in their turn, and thus get further back in time
to the beginning of language. But no one has succeeded
in doing this in a generally accepted way, nor is there
any prospect of it.
The incapacity may be stated in terms of time. The
several outputs of the comparative method take us
back perhaps 6,000 years. (This is the rough time-depth
that Pedersen [1931, p. 319] proposes for Proto-
Indo-European; the estimates reported in Cardona,
Hoenigswald, and Senn [1970] do not go back so far.)
According to any dating that places the origin of lan-
guage appreciably earlier than that, there is an appre-
ciable temporal gap between the origin(s) and the
earliest states that we can reconstruct by the compara-
tive method. According to Ussher's chronology, there
would be a gap of about 1,700 years (since he gave
2347 B.C. as the date of the Babel episode); according
to timetables furnished by the physics and the geology
of the late nineteenth century, there could be at most
a gap of a few thousand years; according to present-day
timetables, the gap might be as great as some hundreds
of thousands of years. (Haas, 1966, discusses repeated
applications of the comparative method; she finds it
possible to repeat the method as much as three times,
but the time depth thereby reached is still only about
five thousand years, i.e., to about 3,000 B.C. See p. 140,
Table 12 and n. 66.)
The fifth point of difference between linguistics and
geology may be summarized as follows. Uniformitarian
geology began with a proposed straightforward method
and a proposed limitation upon its scope. Comparative
linguistics began (at about the same time as Lyell's
geology) with a roughly defined task, and required
about thirty years to work its way through to a more
precise conception of its task and to the formulation
of a method—the comparative method—for achieving
it. The culminating figure was August Schleicher (es-
pecially his work of 1861-62), to whom we next turn.
The founders were Rasmus Rask, Jakob Grimm, Franz
Bopp, and August Pott, whose manner of discovery
has been aptly compared by Antoine Meillet (see
Jespersen [1922], p. 55) with that of Columbus. By the
1860's, as the inherent limitation of the comparative
method became clear, linguists became aware of a veil
like the “veil of stratified rock” which Lyell (according
to Huxley) acknowledged from the outset.
So much for the differences between geology and
paleontology on the one hand, and linguistics on the
other. Let us close this section by noting an important
resemblance. Whitney, quoted above, called the rec-
ognition of the Indo-European family a turning point
in linguistics. The reason why it was a turning point
is revealed by a phrase in the famous remarks (1786)
of Sir William Jones which are generally given credit
for starting the new turn. Greek, Latin, and the newly
discovered Sanskrit, Jones says, show an “affinity...
so strong that no philologer could examine all the three
without believing them to have sprung from some
common source which, perhaps, no longer exists...”
(Jones [1788], pp. 421-22). This prospect was exciting
for two reasons. First, for the thought—not new, as
Hoenigswald (1963, p. 3) notes, but newly interesting—
that a major language should have entirely dis-
appeared. The comparable interest in extinct plants
and animals is well described by J. C. Greene (1959).
In each case, apart from the supposed bearing upon
the revelations of the Bible, there was food for
uniformitarian thought: the proposition that the laws
of nature do not change over time did not mean that
the states of nature do not change; the uniformitarian
did not have to maintain that at all times the species
of plants and animals, and the languages, were just
what they are now. The second reason for excitement
was the hope, and the prospect, that the lost source
language could be recovered. The analogy between the
hoped-for method of recovering the lost language and
the method of recovering extinct plants and animals—
between the comparative method of linguistics and
paleontology—is a rather remote one, but we know
from remarks by Hegel, W. von Humboldt, and others,
that Cuvier's comparative anatomy inspired the foun-
ders of comparative linguistics.
