Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M.

Argüello gArcíA
139

KEYWORDS: Petroglyph – Archaeological excavation – Rock art production – Colombia

ARCHAEOLOGY OF ROCK ART:

A PRELIMINARY REPORT OF ARCHAEOLOGICAL
EXCAVATIONS AT ROCK ART SITES IN COLOMBIA

Pedro María Argüello García

Abstract. This article is a preliminary report of archaeological excavations at some rock art
sites in Colombia. The aim of this research is to obtain contextual information about the pro-
duction and use of petroglyphs in order to overcome one of the principal obstacles for an
archaeological treatment of rock art: chronology. Three kinds of contexts possibly related to
rock art use are examined and discussed.

Introduction considered rock art as a truly archaeological object.

Archaeological approaches to rock art are currently It means that rock art has not been systematically
growing and expanding (Taçon and Chippindale studied or scientiically put into context with other
1998). Diferent ways to understand rock art as an archaeological objects, and, as corollary, it has not
archaeological object have been tested and cumulative been used as a legitimate source of knowledge about
information is nowadays available. However, data pre-Hispanic societies (Argüello 2001). Several argu-
obtained from archaeological excavations, focused on ments, related to Colombian history, as well as social,
contextualising rock art, are yet limited or inexistent academic and scientiic context, would be wielded
(Keyser 2001: 118–9; Bednarik 2007: 120), and their in order to explain the former described situation
potentiality is still overlooked (Whitley 2005: 60). (Muñoz 2006). However, it is a fact that one of the
In Colombia, archaeological approaches of rock art most limiting factors is the lack of reasonable rock
have basically consisted of bringing together rock art art chronologies (see Keyser 2001: 116–7 for similar
and other archaeological objects, but without some urging for U.S. case). It is because of the generalised
methodological considerations; in other words: rock tendency of adopting ad hoc rock art chronologies,
art has been just joined to the rest of archaeological or cultural atributions, but without a scientiically
objects but without providing arguments about the contrasting method or congruent arguments (Pérez
nature of this association. de Barradas 1941; Silva 1961; Duque 1965; O’Neill
The aim of this paper is to present and discuss some 1970; Arango 1974; Correal and Van der Hammen
preliminary results of archaeological research whose 1975: 175; Correal 1979: 233–4; Correal and Pinto 1983;
goal it is to obtain information about production Cardenas-Arroyo 1998). As a result, archaeological
and use of rock art in Colombia. The main starting research found the lack of dating as a wall blocking
point is the idea that several archaeological objects, possibilities for associating rock art as starting point
diferent to rock art itself, could be recovered from for answering questions about the past (e.g. Lleras
archaeological deposits in the areas adjacent to rocks 1989; Langebaek 2001; Bray 2005: 103), condemning it
with petroglyphs to secure useful data for establishing to be just documented (Von Hildebrand 1975; Morales
a contextual approach for dating and explaining rock and Cadavid 1984; Becerra 1990), or, even worse, a
art function. silent companion in archaeological reports (e.g. Casas
1991).1
Rock art and archaeology in Colombia 1 This situation could perhaps explain why atention to
Since its establishment as scientiic practice, ater rock art as archaeological object is inversely proportional
the irst quarter of the twentieth century (Langebaek to the increase of archaeological research. A quantitative
2003; Botero 2006), Colombian archaeology has not analysis about the archaeological literary production
140 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA

1977; Correal 1979; Correal and Pinto 1983). Despite

these findings, which should have encouraged
researchers to secure ages of rock paintings through
analytic methods, they have continued referring
to rock paintings as Muisca-Chibcha without pro-
posing arguments for this atribution (Correal and
Van der Hammen 1977; Correal 1979), and using
ethnohistoric sources about the Muisca-Chibcha as
a base for the explanation of rock paintings.
Another common way to assign chronology to
rock art is by superimposing it on distributions of
datable archaeological objects, named ater Spanish
chronicles (Silva 1961). This typical culture-history
procedure is based upon a normative idea that
homogenises space and time in order to build
cultural areas, puting aside variability produced by
social and political factors. In the same vein, stylistic
analysis, also based upon normative ideas, takes a
minimum part of the rock art images to compare
with datable archaeological objects (Arango 1974;
Correal and Van der Hammen 1977; Correal 1979;
Cifuentes 1989). However these comparisons are so
general that they permit to assign rock art to almost
any culture in the world (e.g. presence-absence of
Figure 1. location of the study area. spirals, circles, triangles). In short, at this time, there
is not any valid dating or chronology for any rock art
The former situation is a consequence of historical site in Colombia.
conformation of archaeological science where each One of the reasons for the lack of consistent rock
object must have entity, space and time, in order to be art chronologies is the absence of scientiic research
understood as an archaeological object. In Colombia, speciically dedicated to dating rock art (Bednarik
every proposed chronology for rock art, based on a 2007). In order to break down the mentioned wall,
variety of approaches such as stylistic analysis, nearest and given that direct dating of rock art is a diicult
association and ethnohistoric analogy would be easily venture (Whitley 2001: 23–4; 2005; Bednarik 2007: 115)
discarded or, at least put seriously in doubt (Argüello — even more in Colombia for the lack of specialised
2001). The main reason is because almost every laboratories, qualiied people and economic limitations
assigned chronology for rock art atributes rock art to on funding — relative dating turns into one of the
indigenous groups identiied by the Spanish ater the most important necessities of rock art in this country.
15th century, based on the premise that the territory However, a date by itself would not be enough or
was occupied only by one group, without temporal useful if it does not bring together contextual maters.
depth (e.g. Triana 1922; Nuñez 1959; Arango 1974). On To airm that a rock art panel is 500 or 5000 years
the contrary, modern archaeology has demonstrated old could be irrelevant, but it is a necessary requisite
not only a deeper occupation but also diferent kinds in order to understand rock art as an archaeological
of societies (hunter-gatherers, tribes, chiefdoms), object (Bednarik 2007: 100). It means that contextual
which renders it necessary to provide arguments for analysis has to be parallel to relative dating.
assigning rock art to some period or kind of society. Among the many ways that have been proposed in
For example, rock paintings of central Colombia, one recent years I consider that understanding rock art as
of the regions where rock art has been reported, was a part of an operative chain would be useful because the
initially assigned to Muisca-Chibcha group (Indians remains of rock art production (Fiore 2007) and use are
identiied by Spaniards in the 16th century) by the able to be archaeologically excavated and associated to
earliest researchers in the 19th century given that, other archaeological objects. Speciically, petroglyph
at that time, no information about chronology was production requires lithic tools whose remains could
available (Uricoechea 1984 [1854]; Restrepo 1972 have been let nearby2. On the other hand, petroglyph
[1895]). However, subsequent modern archaeology
found not just an early occupation (at least 11 000 2 Before the beginning of archaeological excavations, the
research team tried to think about formal characteristics
years), but diferent kinds of societies — some of them
of tools for making rock art on the basis of petroglyphs
setled near rock art (Correal and Van der Hammen themselves (made by indirect percussion) and concluded
from 1800 to 1962 showed that whereas the archaeologi- that they could have been sharp triangular lakes. Neverthe-
cal production was growing, the interest in rock art was less, a parallel experimental program discards this option
decreasing in amount and importance (Jaramillo and and is currently analysing traces in diferent kinds of tools
Oyuela-Caycedo 1994: Table 3). made on several raw materials.
 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA
141

Figure 2. Rock art of El Colegio and some associated elements: mortars and sharpening grooves.

use is to be expected given that used grindstones Muñoz 2006).

have been found in the same rocks, without doubt From the beginning of the 20th century, petroglyphs
associated with them. Furthermore, understanding have been atributed according to the presence of
rock art in technological terms brings the possibility Panche tribes, as a correlate of cultural diference
to engage it in a contextual approach where multiple with their neighbours, the Muisca, who have made
lines of evidence allow recovering information about paintings (Triana 1922). As has been noted, the lack of
diferent aspects related to meaning and function. It chronology allowed irst researchers to think in just a
is necessary to carry out detailed analysis to avoid single occupation or group, so it was natural to assign
simple and doubtful associations that, at the end, petroglyphs to that tribe. Later, O’Neill’s (1973) and
do not contribute to the advance of rock art science Arango’s (1974) archaeological research maintained
(Argüello 2008). In Ramenofsky and Stefen’s (1988) the same idea about a single occupation and assigned
terms, it is necessary to build reliable dating tools in excavated potery to the Panche tribes, reinforcing as well
order to set up valid chronologies. the petroglyph’s assignation by means of similarities
between some designs in potery and petroglyphs
Rock art archaeology in El Colegio and showing differences with paintings through
The western side of East Mountains (Fig. 1) is one stylistic analysis. However, subsequent indings and
of the regions where most rock art sites have been archaeological research showed a very ancient and
reported in Colombia (Duque 1965; O’Neill 1973; diverse occupation (at least 16 000 years old) (Cardale
Arango 1974; Peña 1991; Muñoz 2006). El Colegio 1976; Mendoza and Quiazua 1990; Peña 1991; Van
municipality is known for the large numbers of der Hammen and Correal 2001), and even presence
petroglyphs, cupules and mortars (metates), sometimes of petroglyphs in areas without Panche pottery
appearing together (Argüello 1998; Rodriguez 1998; (Rodriguez et al. 2002). As a consequence, nowadays
Muñoz 2006) (Fig. 2). Nevertheless, the number of at least three archaeological periods are accepted
rock art sites contrasts immensely with the knowledge (Preceramic 16 000 bp – ?, Herrera 2700 bp – 1100 bp,
about them. explicative theories based on questionable and Pubenza 1100 bp – 500 bp) that relativise traditional
cultural ascriptions have been proposed (Triana 1922; petroglyph chronology and relegate it to the status of
O’Neill 1973; Arango 1974) and, by the way, have with a hypothesis, not a truth.
time become true. Likewise, cultural atribution based Archaeological research in El Colegio municipality
on ethnohistorical sources has assigned all the rock art has been planned in order to answer diferent questions
(consisting almost exclusively of petroglyphs) on the related to rock art production and use. Several ield
western side of East Mountains to the Panche tribes, strategies have been tested and adjusted following the
following a standardised cultural-history academic nature of archaeological deposits and indings, types
tradition (Duque 1965; O’Neill 1973; Arango 1974; of sites and particular questions. Due to space, just one
142 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA

Figure 3. Buried rock art.

of the research strategies is presented in this report. It

consists of archaeological excavation at rock margins
aimed at securing information about production and
use. On the other hand, although production and
Figure 4. A buried rock with cupules found together
use would have been related, each is being treated
with two cores.
separately to avoid imposing mistaken relations.
Because analyses are still in progress, only information
especially related to function is discussed. To date, or by coincidence. The most important evidence
twelve rocks with art have been excavated and results against natural concurrence is that buried blocks
could be reported according to the kind of information with petroglyphs do not belong to larger rocks or
recovered: buried rock art, evidence of activities near any others nearby. It means that buried rock art was
to rock art and objects placed in rock art sites. not broken and fell down from the bigger rocks with
petroglyphs or was not transported there by natural
Buried rock art agents. Additionally, transportation is discarded in the
In the bed of two rocks, pieces of rocks with pet- irst case, given that size and weight of the buried rock
roglyphs were recovered3. Although the sample is makes it almost impossible — unless mass movements
too small, there is a correspondence between the not recorded at the site have occurred (Fig. 3).
kind of rock art on the surface and the buried rock In the second case, it is evident that a rock with
art. In other words, formal images and cupules were cupules was placed together with two cores just at
buried where formal images and cupules were made. the larger rock edge (Fig. 4). Due to the slope being
The main question is whether buried rock art was oriented in the opposite direction to which objects
intentionally placed or just occurred there randomly were excavated, natural and simultaneous occurrence
of them as a consequence of soil movement is unlikely.
3 It is not the irst time that buried rock art is reported in If small rocks with cupules and cores have been
Colombia. Previous researches by German Peña (1988) in
transported by gravity — from an upper zone of the
Cachipay municipality (approx. 15 km from El Colegio)
site — they would have been deposited on the top of
found a buried rock with cupules, although not on the
bed of rock art. However, given that the aim of this study the main rock, not on the botom where they were
was not related to rock art, the researcher did not propose recovered. However, anthropic intervention is still
any kind of association between rock art, potery shards a possibility. Human occupation subsequent to pre-
and lithic tools excavated together. Hispanic times has been documented in the site but the
 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA
143

Figure 5. Burnt area between two rocks with cupules.

exact nature of its seting

is not yet understood.

Activities near the rock

art
Two cases of activi-
ties near two rock art sites
have been documented.
Both of them are related
to burning activities.
The irst case comprises
burnt soil between two
rocks with cupules.
At the junction of the
rocks, a burnt area
associated with potery
shards, chipped lithics
and a post mould was
found (Fig. 5), although
secure association
of burning activities
with rock art remain
problematic because
the site where rock art
is placed corresponds
to a possible household
area as well (Fig. 6). To
know if the burnt area
is related to domestic
activity or to the rock
art is very important, Figure 6. Possible household area (correlated by shard distribution) near rock art (see Fig.
not for creating a dicho- 5).
tomy, but for helping
accuracy in rock art dating. Relation between domes- three ground depressions (mortars) on the main rock
tic activities and rock art production and use could and associated to cupules and other petroglyphs (Fig.
be explored once activities related to rock art have 7). In this case, the inclination grade of the slope in
been proved. the area where the rock art is placed would not have
The second case is the discovery of some rocks allowed the seting of a household, or to perform some
that were used in a irebox together with rounded large-scale activity, so it is possible that the irebox
cobbles (cantos rodados), one macerator, potery shards was associated to the activities performed on the rock
and chipped lithics. No burnt soil was found in the with the rock art.
site, but buried rocks show evidence of ire exposure
(soot spots and issures produced by exposition to Objects placed at rock art sites
high temperature). The irebox was set in a drain of In nine of twelve excavated sites archaeological
144 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA

Figure 8. Distribution of cobbles from 36 excavations

at (a) rock art margins and (b) away from rock art
margins.

chaeological excavations have been made in nearby

sites lacking rock art, it is possible to evaluate the
occurrence of cobbles comparing their frequency at the
sites with and without rock art. In fact, some test pits
and archaeological excavations not in the immediate
Figure 7. Location of a irebox in a drain of three vicinity of rock art have occasionally yielded one or
mortars. two cobbles, while excavations on the margins of

objects have been recovered. Findings

consist on potery shards, chipped lithics
(artefacts and waste products), polished
lithics and non-modiied lithics. The nature
of these artefacts is currently being studied.
Here only one kind of object is considered:
cobbles without modiication evidence whose
non-functional character and frequency of
appearing could refer to activities related
to using large rocks with petroglyphs.
Rounded cobbles did not occur naturally,
so they were brought to the sites that
were excavated. Additionally, they appear
generally near the rock art. Given that ar-

Figure 9. (a) cavity naturally formed, and (b)

adjacent area where (c) an unusual number
of cobbles was found.
 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA
145
decorated rocks have produced up to six (Fig. 8).
One case deserves special mention. In a rock with
cupules there is a small cavity formed naturally by two
rocks where six rounded cobbles, potery shards and
chipped lithics were deposited. Given that the cavity
is closed in the back and top, and the only access is
opposite to the slope direction, the only possibility
is that the objects were placed there intentionally by
somebody, not naturally. Moreover, in the adjacent
area to the cavity a concentration of archaeological
remains was noted, so it is also a possibility that they
had come from it and been transported by natural
processes (Fig. 9).

Preliminary results
The most important preliminary result of archaeo-
logical research in El Colegio, regarding rock art
chronology, is that it has rejected the traditional
theory that the petroglyphs were made by Panche
tribes (Urbina and Duarte 1989; Muñoz 2006). To
date, potery that would be atributed to Panche has
not been found in any of the previous archaeological
projects (Rodriguez et. al. 2002; Rodriguez 2005), in
areas near to rock art. Potery found in test pits and
archaeological excavations, close to or in the beds of
petroglyphs, belongs to the Herrera and, possibly,
early Pubenza Period (Fig. 10). It means the zone was
setled at least 1500 years before previously thought, Figure 10. chronology of el colegio (according to
and late groups, presumably Panches, apparently potery) vs previously claimed rock art chronology.
did not live there.
Given that there is not much evidence of Panche refers to portable pieces of rock with petroglyphs
occupation it is inadequate or not very useful to deposited on the beds of main rocks with petroglyphs;
continue using ethnohistoric references as a source so dating refers more to use than production. Burial
for explaining rock art, in particular, and pre-Hispanic of rock art by natural factors is unlikely and the way a
setled group’s life in general. Previous archaeological small rock with cupules was placed together with two
research has showed that, although Herrera cores suggests intentionality. This idea is reinforced
populations were antecessors of later populations, by indings of cobbles located in a cavity. Both buried
they were diferent groups — in sociopolitical and
economical terms (see Langebaek 1995; Boada 2007); so
ethnohistoric sources are not valid for understanding
El Colegio pre-Hispanic occupation, unless carefully
and argued statements are presented.
The occurrence of lithic tools in the beds of
rocks with petroglyphs is a hopeful ind in order to
understand rock art production. Even more, lithics
found together with potery in the same layer opens
the possibility of dating such events. The presence
of diferent kinds of chipped lithics, from diferent
steps of production sequences, reveals which tools
manufacture was carried out in those places (Fig.
11). An experimental program has not yet achieved
positive identiication of petroglyph-making tools, but
recovered lithics have allowed securing preliminary
information such as raw materials, tool production Figure 11. chipped and polished lithics from 25
techniques and inal artefacts. excavations (1 � 1 m) at rock art margins (�F
Buried rock art has been one of the most useful – tabular fragment, Uc – used core, NUc – not
tools for rock art dating, although such atempts are used core, UF – used lake, NUF – not used lake,
usually related to in situ rock art covered by sediment DEB – debitage, PS – polishing stone, PEF – pestle
(Bednarik 2007: 120–1). In El Colegio, buried rock art fragment).
146 Rock Art Research 2009 - Volume 26, Number 2, pp. 139-164. P. M. Argüello gArcíA

rock art and cobbles could have been offerings or of this phenomenon. Archaeological objects ofer the
payments, practices that have been documented among opportunity for controlling hypothesis about meaning
contemporary Indians in Colombia (Javier Rodríguez, and function and allow the defeat of ad hoc and
pers. comm. 1998). However, dating of the buried rock comfortable theories whose aim is to explain rock art
art is elusive because no potery was found together easily. On the other hand, an archaeological approach
with it. would contribute to melting pessimistic ideas about
Potery shards placed together with cobbles in the chance of scientiic understanding of rock art.
a cavity render a possible dating for these kinds of Maybe these eforts could in the near future refute the
activities viable. Although not all potery shards in premise by Vicente Restrepo (1972 [1895]: 212) at the
the cavity have been positively identiied, it is clear end of 19th century: Colombian rock art ‘…can reveal
that at least some of them are of the Herrera Period. nothing to science’.
There is no conclusive evidence that cobbles and
related activities were associated with petroglyphs; Acknowledgements
archaeological excavations merely inform us which Juan Carlos Rodríguez is co-researcher of this project.
rock was used since 2700 years ago. Without the help of Rodriguez-Buitrago’s family, especially
Finally, archaeological excavations have conirmed Javier, Elsita and Luis Eduardo, this research would never
have been made. Alcaldía Cívica de El Colegio provided
that diferent kinds of activities took place near and in
economical funds for ield sustenance. The Students of
the rock art sites. This idea was anticipated by general Anthropology Program at Universidad de Caldas took part
finds of mortars, work surfaces and sharpening at ield seasons and laboratory analysis. Ramiro Bejarano
grooves on the same rocks as petroglyphs, even and José Lloreda kindly gave authorisation to carry out
sometimes forming apparent compositions. It is not archaeological excavations in their properties. Francis
clear if such activities were performed before, parallel Paola Niño patiently helped with English translation. Two
to or ater petroglyph production. Potery analysis rAr reviewers made pivotal comments that contributed to
airms that some of such events took part during the reining earlier arguments. I thank all of them.
Herrera Period, and spatial analysis has shown that
Pedro María Argüello García
some of those sites were used for domestic purposes
Departamento de Antropología y Sociología
too. Universidad de Caldas
Manizales
Colombia
Conclusion
E-mail: [email protected]
Finding of archaeological objects below rock art
panels would open the door for beter understanding Final MS received 21 February 2009.
RAR 26-922
of many aspects of rock art production and use.
Archaeological excavations of rock art sites in El
Colegio have brought the possibility of a promising
approach to relative dating of petroglyphs in Colom-
bia. Likewise, preliminary conirmation of the use
of rock art sites extends the range of archaeological
objects that could be recovered and, by the way, the
opportunity to know a litle more about the context
in which rock art was involved. At least three kinds COMMENTS
of objects and contexts, possibly associated to rock
art use, have been found and systematic future About archaeology and rock art
excavations could yield more.
Nonetheless, the recovery of archaeological objects By ROY QUEREJAZU LEWIS
on the margins of rocks with rock art should not,
by itself, open the door for simplistic association These comments, based on the paper by Pedro
(e.g. Mendez 2008): proximity does not necessarily María Argüello, are intended to lead us to ask our-
mean association. Information from archaeological selves about the validity of archaeological methods
deposits must be carefully evaluated and systematic in rock art research.
research needs to be followed by discovering paterns Argüello’s paper is fundamentally based on archae-
in the evidence. Isolated indings are a suggestive ological methods aimed to obtain reliable information
source for new questions about production and on the context in which the referred rock art was
use of rock art, but careful analysis of each context produced and when it was produced. It shows very
and its comparison with cumulative facts remain a clearly that rock art research in Colombia is taking its
methodological requirement. This means at least three irst steps, which most rock art research trajectories
scales of analysis: site, locality and region, in order to worldwide have passed. At the same time, Argüello’s
contrasting the reliability of indings. paper is representative of a great part of rock art
Eforts for understanding rock art as an archaeo- research in South America. This is understandable,
logical object are the irst step for a beter comprehension mostly because research is starting from zero under