ay practices Empiricism and model building in stratigraphy:
The historical roots of present-day practices
Andrew D. Miall
Department of Geology, University of Toronto, Toronto, Ontario M5S 3B1, Canada
email: miall@geology.utoronto.ca
ABSTRACT: The science of Stratigraphy has, since its inception in the late eighteenth century, been characterized by two contrasting
research modes or cognitive styles (Rudwick 1982). Empirical (inductive) descriptive stratigraphy began with William Smith, led to
the establishment of a data base of stratigraphic units (Murchison, Sedgwick, Lapworth), and formed the basis for modern work to es-
tablish and refine a detailed chronostratigraphic time scale (Van Hinte, Berggren). Other workers (Hutton, Lyell, Darwin, Chamberlin,
Ulrich, Umbgrove, Sloss, Vail) have sought to identify underlying geological controls, and have built deductive models to explain earth
processes, beginning with Huttons uniformitarianism. Many such models sought evidence of regularity or cyclicity in earth processes
(the pulse of the earth), including the modern global-eustasy model of Vail.
There is an ever present danger that models can drive the analysis and presentation of data, particularly where stratigraphic models
have been invoked to explain, clarify or codify the stratigraphic record. These problems are not new. Attempts to apply European
chronostratigraphic units to North American stratigraphy in the early twentieth century were accompanied by expectations that unit
boundaries would be marked by lithologic events, such as unconformities. These expectations were not supported, and this may have
been the basis for North American attempts to establish alternative stratigraphies, including what became sequence stratigraphy. Ulrich
(1911) thought that stratigraphic successions were created by diastrophic cycles, and was concerned that regional correlations of these
successions did not appear to be supported by the biostratigraphic evidence. Barrell (1917) was one of the first to understand the prob-
lems created by the lack of representation of long intervals of time in the geologic record, and developed ideas concerning the relation-
ship between base level change and sedimentation that we now term accommodation.
Modern work on the chronostratigraphic time scale is based on empirical principles, culminating in the definition of global section
and boundary stratotypes for the major chronostratigraphic units. However, a controversy has recently arisen over the preference by
some geologists to use distinctive marker events to define boundaries. In some cases, this involves introducing hypotheses about the
global extent and geological superiority of such events, rather than relying on the accumulated historical record of biostratigraphic and
other data.
INTRODUCTION
Geology is historically an empirical science, firmly based on
field data. Hallam (1989, p. 221) has claimed that Geologists
tend to be staunchly empirical in their approach, to respect care-
ful observation and distrust broad generalization; they are too
well aware of natures complexity. However, interpretive
models, including the modern trend towards numerical model-
ing, have become increasingly important in recent years.
The empirical approach to geology, including the building of
models, is inductive science, whereas the use of a model to
guide further research is to employ the deductive approach.
This methodological difference was clearly spelled out for ge-
ologists by Johnson (1933). Frodeman (1995) recently re-
viewed the work of the German philosopher Heidegger, arguing
that the practice of the science of geology illustrates a process
termed the hermeneutic circle, in which induction and deduc-
tion supposedly follow each other in an iterative process of ob-
servation, generalization and theorizing (induction), followed
by the construction of hypotheses and the seeking of new obser-
vations in order to test and abandon or refine the theory (deduc-
tion). Ideally, this is a continuous and circular process (text-fig.
1), but we have argued elsewhere (A. D. Miall and C. E. Miall,
this volume, in a paper that expands on the implications of her-
meneutics for stratigraphy) that at the present day there are sep-
arate groups of stratigraphic researchers that are separately
following these two different methodological approaches in
partial isolation from each other. The purpose of the present pa-
per is to argue that this dichotomy has deep historical roots; that
from the mid-nineteenth century to the present, the inductive
and deductive approaches to the science of stratigraphy have
largely been followed by different groups of researchers having
different objectives, and that throughout much of the history of
the science, the groups have had little to do with each other.
Since modern stratigraphic studies began in the late eighteenth
century a central theme of stratigraphic research has been the
empirical construction of a vast data base of descriptive stratig-
raphy, focusing on the occurrence and relative arrangements of
formations and their contained fossils. This data base now con-
stitutes what has come to be called the chronostratigraphic time
scale. In recent years, methods of determining the ages of beds
by other means, such as by radiometric dating, magneto-
stratigraphy and chemostratigraphy have added depth to this
data base. As we show here, research into the preserved record
of deep geological time has grown into an enormously complex,
largely inductive science carried out mainly in the academic
realm. From this, a descriptive (inductive) classification of
Earth history has been built, consisting of the standard eras, pe-
riods, epochs, etc.
At various times deductive models of Earth history have been
proposed that have had varying levels of success in contributing
stratigraphy, vol. 1, no. 1, pp. 3-25, text-figures 1-8, 2004
3 to our understanding of the Earths evolution. There have also
been many attempts to develop deductive models of strati-
graphic processes, including the cyclothem model of the 1930s,
and modern facies models and sequence stratigraphy. Ideas
about the tectonic setting of sedimentary basins have also in-
cluded several bold attempts at model building, including the
pre-plate tectonic geosyncline theory of Kay (1951), the mod-
ern petrotectonic assemblage concepts of Dickinson (1980,
1981), and the various geophysically-based basin models of
McKenzie (1978), Beaumont (1981) and many later workers.
Some of the concepts of sequence stratigraphy, that evolved
from seismic-stratigraphy in the 1970s, constitute one of the
most recent and most elaborate attempts to develop deductive
stratigraphic models. These included the Exxon global cycle
chart (Vail et al. 1977; Haq et al. 1987, 1988), which, if it had
proved to be a successful explanation of the stratigraphic re-
cord, could potentially have become the dominant paradigm,
entirely replacing the old inductive classification of geologic
time, and largely supplanting the complex set of methodologies
with which it was being constructed. As we have discussed
elsewhere (A. D. Miall and C. E. Miall 2001; C. E. Miall and A.
D. Miall 2002), the two distinct intellectual approaches resulted
in the development of two conflicting and competing para-
digms which are currently vying for the attention of practicing
earth scientists. In a companion paper (A. D. Miall and C. E.
Miall, this volume) we argue that current research in the field of
cyclostratigraphy may be following a similar pattern of devel-
opment
The history of stratigraphy since the end of the eighteenth cen-
tury has encompassed the following broad themes:
1) Recognition of the concept of stratigraphic order and its re-
lationship to Earth history, and the growth from this of an em-
pirical, descriptive stratigraphy based on sedimentary rocks and
their contained fossils.
2) The emergence of the concept of facies based on the rec-
ognition that rocks may vary in character from place to place
depending on depositional processes and environments. This
was one of the first deductive models developed to facilitate
geological interpretation.
3) Recognition that rocks are not necessarily an accurate or
complete record of geologic time, because of facies changes
and missing section, and the erection of separate units for
time and for rocks.
4) Development of a multidisciplinary, empirical approach to
the measurement and documentation of geologic time, an unfin-
ished science still actively being pursued to the present day.
5) Attempts at different times to recognize patterns and themes
in the stratigraphic succession and to interpret Earth processes
from such patterns. Facies models and sequence stratigraphy
are amongst the main products of this effort.
6)Attempts to extract regional or global signals from the strati-
graphic record and to use them to build an alternative measure
of geologic time, based on an assumed pulse of the Earth.
These themes may be further generalized into a descriptive, in-
ductive approach to the science (themes 1 to 4), which we here
categorize as the empirical paradigm of stratigraphy, and a dis-
tinctly different, interpretive approach to the subject (themes 5
and 6), that we term the model-building paradigm. Elsewhere,
we have examined the model of global eustasy as applied to se-
quence stratigraphy (A. D. Miall and C. E. Miall, 2001, C. E.
Miall and A. D. Miall, 2002), the use of climate proxies in
stratigraphic studies (A. D. Miall and C. E. Miall in prep) and
the use of cyc