 |
ARKAMANI Sudan Electronic Journal of Archaeology and Anthropology |
February 2005
POTTERY DECORATION IN PREHISTORIC
A NEW PERSPECTIVE
By Isabella Caneva
Barbara E. Barich (ed.) 1987, Archaeology and Environment in the Libyan Sahara: The excavations in Tadrart Acacus, 1978-1983, Cambridge Monographs in African Archaeology 23, BAR International Series 368: pp.231-254.
The problem of classifying prehistoric archaeological
materials in the
The need of more significant typologies has led the author to emphasize technological procedures rather than morphological aspects in pottery classification. Technological procedures, indeed, represent the synthesis of the operator's investment in terms of labour, time, care and taste.
In the central
A comparable trend can be seen in the chronological development of Saharan pottery. A certain parallelism can be followed especially in the shift from the rocker comb impression to the compass-like application of a two-toothed instrument. This shift is particularly emphasized in both areas by the evolution of the "dotted wavy line" decoration, obtained first with the rocker technique and later with the alternately pivoting stamp.
As far as Uan Muhuggiag
is concerned, its pottery appears to be decorated with techniques belonging to
the later phase of the a forementioned trend.
The most common is the "return" technique, which consists in the compass-like
application of a two-toothed tool, going over the last line of dots, thus
ensuring equal spacing between all the dots of the pattern. This results to be
the most common technique of decoration all over the
Despite similarities and parallelism, substantial dissimilarities are also recognizable between the Saharan and the Nilotic areas. Many elements point, therefore, to a normal framework of exchange and contacts between adjacent regions, sometimes influencing the local production, but never in terms of predominance of one region over the other.
The elaboration of true and proper typological structures,
which are
When dealing with ceramics the precise aims of classif
a) To differentiate the various ceramic elements, reconstruct the complexity of the ceramic production and assess chronological variations.
b) To recognize significant elements which permit easy comparisons among contexts.
c) To have a clear view of the amount of labour involved in
the
production, the time required, the special care needed,
d) To show the importance of certain activities over others. This is normally related to the shapes of the pots and, therefore is not considered relevant in contexts where shapes can not be reconstructed.
As may be seen from this very short general list of goals, the basic premise in selecting typological elements (or associations of elements) is that the choice must come after observing the material and formulating the specific research aims, and not decided upon in advance.
In any case, the typological elements must have the following properties:
a)
b)
c)
It is noted that these requisites are difficult simultaneously in a single element; they are more often founc several elements either together or at different levels of s case, the structural order of the headings (division into m sub-groups) can be decisive for either grouping the data ai significant, or dispersing them, even if in theory it should read the list in different ways.
In prehistoric Saharan ceramics the available elements a
the one hand, by the poor state of conservation of the materi it impossible to reconstruct
complete shapes or complex decoz on the surface of pots and, on the other hand,
by the great both the make up of the ceramic paste and the surface treatme
The only analysable element without laboratory aid is,
decoration. There are, however, different ways in which the be evaluated,
namely:
a) the instrument used;
b) the decorative patterns achieved;
c) the technique adopted to produce the
decoration.
It is difficult to
separate these aspects, one from the other. The
visual impression
obviously takes all three into account. However,
the rationalization of the visual impression must give them a sequence. These
sequences must follow a logical pattern which is not necessarily either the
succession in time of the various acts performed or the content of the
"visual effect" (in this sense, as already mentioned, the elements
are inseparable), but reasoning based on the quantity of information which each
element possesses.
These three aspects of the decoration were adopted for the
first Saharan ceramic typology (Camps-Fabrer 1966; Camps 1968), and then,
conventionally, by all archaeologists who subsequently dealt with the subject,
virtually without further questioning. The
sequence in which these aspects were considered is usually that set out above,
whereas in classifying Sudanese ceramics it is the motifs that are considered
first (Arkell 1949; 1953; Haaland 1981; Hays 1974; Mohammed Ali 1982). However, the author feels that the approach
calls for two remarks: the first is general and theoretical and concerns the
order followed, the second is practical and particular and concerns the
analysis of decorative techniques.
1. THEORETICAL COMMENT
The classification of
decorative motifs is not based on syntactic arrangement, which can only be
reconstructed in extremely rare cases, but on the optical impression given by
the individual decorative elements. This,
by the way, is not easy to describe and is extremely subjective both for the
author and for the reader of the description. The grouping into
"classes" is particularly subjective (the motifs often overlap
classes), so that there is a risk of either having groups which are too
extensive and generic (dotted patterns, wolf's tooth, etc.) or going to the
other extreme and separating almost every fragment. In
all probability, the craftsman was subjective in using motifs generally
accepted by the community and varying his execution of it from one time to the
next.
The same may be said for the implements used, for which
there were no standardized materials, shape or size permitting them to be
associated specifically with decoration motifs. Tools were probably made or
collected individually and were, therefore, merely personal objects. Therefore, this classification is no
longer based on real objects present, but on objects whose existence and
physical aspect must be assumed. This
procedure, then, although necessary for formulating assumptions about
descriptive details, does not attain the value of a useful classification,
since it becomes as subjective as the decorative motifs.
By considering the technique used before taking the other
elements into account, the following advantages are achieved:
1) the technique is a fact; and
does not lend itself to individual subjective interpretation; it can be
recognized in the same way by anybody and is easily described;
2) it provides a great deal of information,
giving an immediate realistic reconstruction of a moment in the past, with
indications as to the types of gestures made, the position of the hands, the
way the pot was held, the care taken in executing the operation, and the
duration and cost of the latter;
3) the technique is very distinctive even with
respect to the decorative motif which is made by it and is not detached from
it. On the contrary, often the
motif can only be done in one particular way (the smocking pattern for
instance). The technique is probably
the first thing that an old craftsman taught a young apprentice, together with
all the motifs it can produce and the different ways to enrich and vary them.
Each craftsman then developed his personal version of the patterns;
4) the technique
reconstructs a "general trend" (a specific way of treating the
material), which conveys one organized meaning to the variability of the motifs
present. It, therefore, goes beyond
a detailed attribute list of myriad of different executions, by accomplishing
an operation which is a synthesis of what is simply seen (1).
As a result of the preceding remarks, which are applicable
to many categories of archaeological material, there is a dichotomy as to the
different kind of information conveyed by a) the technical and b) the
morphological aspects of the materials. It
has been decided, therefore, to adopt for Saharan and Sudanese prehistoric
ceramics a criterion for classification based on the techniques used and then
make internal subdivisions which take into account the other aspects
mentioned, in the following order:
- Technique of decoration
- Type of instrument used
- Resulting decorative morphology
This method, applied to Sudanese prehistoric ceramics
(Caneva 1983, 1984) yielded highly satisfactory results both for the
reconstruction of a sequence and for the internal differentiation of cultural
aspects. Chronologically speaking, this made it possible to identify a
technological trend which relates all cultural phases from the early Mesolithic
to the late Neolithic, a trend which appears to follow similar phases also in
the
The importance of thc technique used for ceramic
decoration compared to the specific shape of each instrument is stressed by
Camps (1968:127), even though the type of intrument then influences his
typology far more than the technique.
As previously pointed out, even in a classification of
motifs considering technical aspects in the internal subdivision, it would have
been possible to observe the differentiation between the two sites and,
similarly, although with greater difficulty due to the dispersion of data, it
would have been possible to reconstruct the trend supporting the entire
development of ceramic decoration in the Nile Valley. However, since the
techniques were the more distinguishing elements, it was logical to give
preference to a criterion of classification which emphasized them more than the
other elements.
2. PRACTICAL OBSERVATIONS
In previous studies of
the decorative techniques used, analysis was mainly carried out without
considering the logic behind the information visually recorded. One of the most surprising cases is
that of the wellknown dotted wavy line decoration: it was imagined that a
half-moon shaped toothed stick was used, directly applied to the pot,
alternating the position of the curve upwards and downwards (Camps-Fabrer 1966:
446). This mechanical movement was repeated on the entire surface of the pot,
for thousands of impressions arranged in hundreds of rows all perfectly
parallel. Assuming this to be so, one
has to marvel at the dexterity of the craftsman who operated in such a way as
to make it impossible to find any trace of joining the juxtaposed impressions
and at the same time obtaining perfectly parallel lines.
°In most cases, however, the curves are not repeated in
exactly the same shape all along the line, nor is the dotting of the different
curves forming the strip uniform, but it is less dense in the external, convex
parts of the hand, and both these features are unexplainable in a repetition of
impressions using the same instrument. The lines are then normally grouped in
bands of 5 or 6, as are the straight dotted lines that usually form the entire
decorative pattern above and below the undulated bands. Finally, an account must be taken of the absurdities of the
suggested procedure in terms of manual dexterity and time required for the
levels of precision achieved. What was proposed was a procedure which was not
suggested by any real element; the individual elements themselves, i.e., dots
and curved lines, were abstracted from the context in which they repeatedly
recur. If, on the contrary, one
succeeds in showing the relationship between these elements (parallelism even
in assymmetry, grouping of lines, position, number and spacing of each dot with
respect to the dots of the lines above and below, frequency of impressions in
modules or panels, etc.) and endeavors to give a logical explanation of the
entire operation in terms of time and ease of execution, then one will obtain a
certain number of elements which will be rational and provide a significance
for what is seen.
This type of observations, together with a careful
analysis of the material from the foregoing point of view, leads to proposing a
procedure which has so far been confirmed (or, at least, not contradicted) in
all cases examined in both Sudanese and Saharan contexts. This consists of
making fans of dotted impressions with a rocker movement of a toothed comb
(fig. 9.1,_5). With this method the dots on the lines closest to the point of
the fan are the closest together whereas on the outer lines they are more
spaced, as in all rocker patterns. This
produces perfectly parallel patterns, accounts for the difference between one
curve and another on the same line and ensures a perfect repetition on
successive lines. It also
accounts for the grouping of lines in bands, wavy or otherwise and finally
for the more widely spacing of the dots in the arc
farthest from the centre Laboratory tests using this system gave good results
in a surprisingly short time (fig. 9.1,2) (2).
A similar case, which arose from the observation of the
Acacus material and was resolved following the same perspective of showing the
"relationship" between the various elements, concerns the direct
application of combs (plain, "filetes", etc.) in parallel horizontal
ra.s. Although the existence of this technique is totally logical, widely
observed in Sudanese contexts and probably to be assumed, in certain cases,
also in Saharan contexts, its recognition must be supported by a very specific
series of relationships between the elements. Basically speaking, the following
characteristics must be encountered in the decoration (fig. 9.2,1):
a) presence of a "module" with
specific characteristics (for instance irregular spacing and shapes), repeated
alone along the line and in the various rows;
b) presence of interruptions, even if well
joined, between one impression and another along the line;
c) minimum spacing of the holes along the
line;
d) irregular (unavoidable) spacing and
rows not perfectly parallel.
If these characteristics are not encountered, there is
good reason to doubt that the decoration considered was executed with a comb
normally impressed in the clay. If this is the case, then one should look for
the following characteristics:
a) unbroken lines;
b) impressions with varied spacings in the
row;
c) rows with impression of different
shapes and sizes;
d) equal distances between perfectly
parallel rows;
Good evidence of the validity of our proposal were also
found in published materials: for the less densely dotted convex waves cf.:
Caneva 1983: fig. 15,3; Arkell 1949: pl. 79; Addison 1949: pl. XCIV,C, 1, 4;
Bailloud 1969: fig. 2,3,15; Courtin 1969: fig. 21, 5, fig. 25; Gabriel 1981:
photo Ic,d; Roset 1983: fig. 5,2,4; Hugot 1963: pl. XVIII a,h, pl. XIX 4,2;
Camps 1968: pl. XVI, 1-3, pl. XVIII, 5; Barich 1974: fig. 30, 1, fig. 33,3,7;
Arkell 1949: pl. 102.1. For the series of fan impressions cf.: Bailloud 1969:
fig. 2,A (last row); Courtin 1969: fig. 25; Gabriel 1981: photo Ic; Roset 1983:
fig. 5,2,4; Barich 1974; fig. 33,6,8. For repeated rows of variously shaped
waves cf.: Caneva 1983: fig. 15,2; Roset 1983: fig. 5,4; Barich 1974: fig.
30,9. For clear evidence of rocker comb impressions cf.: Caneva 1983: fig. 15,
5-7, fig. 13,4; Arkell 1949: pl. 72,5; Bailloud 1969: fig. 2,A.
e) impression placed regularly in an
intermediate position with resFezt to the row above and below;
f) spaced rounded tooth impressions.
If some of the above characteristics are present (a, b, c,
d), the pattern was probably made vertically, with a comb used pivot fashion
(3) (fig. 9.2,4-5). If all are present, they were probably made with
double pointed tools using compass-type movements, going over the holes already
made for each new pair of rows (fig. 9.2,2). However, if d) is not present then
the same technique has been used but without going over the holes =. second
time (fig. 9.2,3).
Another case which is frequently misinterpreted is that of
a decoration formed by a series of opposed triangles or 'Y's (fig. 9.3,3). Very
often, in both Saharan and Sudanese contexts, it is considered as a series of
normal impressions made with a triangular-section stick (or quoin), in the
opposite direction in every row (Haaland 1981; Roset 1978). It must be noted,
however, that this type of impression, which in some cases recurs, gives a very
different effect, due to the inclined plane of impression and to the less acute
angle (fig. 9.6,1).
Arkell already pointed out the symmetrical nature of the
pattern and recognized in it the effect of a double-pointed instrument, making
impressions in pairs of lines (1953: 71).
In the
A good example is the
fragment from Amekni (Camps 1968: p
l. XIX, 1) with two and a half bands of
rocker decoration on 12 rows (5 with simple oblique dots, 2 with double dots - on
the middle of the pattern, where the dots are gone over twice, thus doubling
the number of impressions - 4 with simple thin horizontal dots) corresponding
to the 12 teeth of a comb, maybe filete, applied vertically with
a pivoting impression. The
same pattern is visible in the other fragment illustrated in the same plate
(No. 4) showing a rocker pattern on 5 rows (two bands of which are
represented): 2 rows of rounded dots at the beginning and end of the pattern, 2
inner rows of bigger dots most of which have been gone over and so close that
they are almost completely amalgamated. We
cannot believe that 5 different combs were used to make this pattern. Moreover, this appears as a typical
rocker pattern, with twice as many dots in the middle rows as there are in the
outer rows, where the comb pivots and, therefore, makes only one dot for each
two zig zags.
of parallel lines, equidistant impressions, matching
positions of the latter (which are always intermediate with respect to those of
the other row), regular repetition, on alternate rows, of different impressions
corresponding to the different shapes of the points. By going over the holes
with the point of a compass with a fixed opening, evidence of the extremely
regular effect (never attainable by simple punching) was obtained.
3. POTTERY DECORATION AT UAN MUHUGGIAG:
DESCRIPTION OF THE TECHNIQUES
Except for the plain
rocker, the pottery of Uan Muhuggiag showed at first sight two contrasting
features: great general homogeneity and different details on every fragment. It, therefore, appeared almost impossible to
build up a typology other than that based on recognition of the use of various
pointed sticks impressed in the paste at various angles, both features being of
such a personal nature that they were unable to provide any significant
grouping.
The attempt to recognize clusters or particular schemes
from the uniform pattern of punctation, was, in most cases, very frustrating:
all the impressions were, without any apparent reason, very regularly spaced.
However, the extreme regularity itself appeared suspicious, suggesting that
very special techniques and tools were used, to warrant both the equidistance
and the "smocking" distribution of the impressions. The rocker
technique could never produce this effect, even when the teeth of the comb are
regularly spaced and shaped, because the vertical zigzag forms horizontal lines
with differently spaced dots. As far
as the normal impression of combs or single points is concerned, this technique
can only accidentally produce the pattern, but it is not a guarantee. The solution was offered by rare cases of
inaccuracy or errors, which suggested that double pronged instruments were used
as compasses pivoting alternately on the points, and in every new pair of lines
going over the dots in the last line (figs. 9.5,6). This way both the perfect spacing and the endless exact reproduction
of the scheme were warranted.
Therefore, from a technical point of view, the two main
groups which were first evident in the ceramics from Uan Muhuggiag (rocker and
punctations) could be further divided as follows:
Plain rocker
We call rocker what in the french literature is called
"impression pivotante". In this case, the tool used is not a toothed
comb, but the plain, thin edge of a spatula or a bivalve shell, applied to the
surface of unfired pots and moved zig-zag fashion, lifting first one end of the
edge and then the other, alternatively. The result is a continuous line, like
an incision, forming various kinds of zig-zag, according to their spacing, as
well as to the form and inclination of the tool. The holder of the tool the direction of the movement, so that the
impression is deep on one side (where the movement is arrested) and side, from
where the tool was dragged.
inclined it towards normally sharp and shaded on the other
At Uan Muhuggiag tools with particularly thin sharp edges
seem to have been used, applied in two very different ways:
a) horizontal bands of curved, lightly
impressed zig-zags. They
are often employed for rim or lip decoration and they represent the most common
application (fig. 9.3,2);
b) vertical bands of zig-zags in which the
tool is so deeply impressed in the paste that it raises backwards a great mass
of clay. This decoration occurs very
often on certain kinds of everted rims (fig. 9.3,1).
"Fish Scale"
It is so defined by Arkell (1953: 73) as a particular form
of plain rocker decoration, in which no space is left between the bands, the
edge of the tool always pivoting on the bottom points of the already made band.
The effect is a lozenge network, which recalls a sort of
net slinging which is still in use in Africa to carry vessels (Camps-Fabrer
1966: 466).
Rare in the Sudan, it is well known in the Sahara, where
it was called "decor guilloche" (Arambourg, Balout 1952: 292).
The applications are the same as in the foregoing pattern,
but they are usually found as body decoration rather than rim or lip
decoration.
Alternate pivoting stamp
This technique is a kind of extreme simplification of the
older rocker technique: a double pronged object is applied to the surface of
the pot and moved horizontally, pivoting alternately on one point and then on
the other.
The result is a pattern of pairs of dotted lines. In each
pair the dots of one line are equidistant and intermediate from the dots of the
other line, which only happens by chance between lines of different pairs.
The dots have various forms, according to the shape of the
tool. Often the points of the same tool differ one from the other, providing a
lively effect of horizontal alternation of different lines (fig. 9.4, _2) The
motifs vary also depending on the way of pressing and moving the tool,
sometimes with lateral dragging in the direction of the movement, sometimes
with greater pressure on one point. The latter produces the impression of the lesser
tapered part of the point, giving the dot a triangular form, as at the top of a
zig-zag design. Very rarely in this context, but frequently in the middle Nile
Valley and also in parts of the Sahara, for instance in the Niger (Roset 1983:
fig. 7), both points give the impression of opposed triangles or vees (fig.
9.3,3).
An effect of horizontal pattern is typical of this
technique, and the closer the dots are on the line, with respect to the space between
the lines, the greater is the effect (fig. 9.2,3). In some cases the pattern
looks as if it was made with a normal comb. However, the difference between
these two techniques is normally recognizable, since a comb is supposed to have
small, quadrangular, closely spaced teeth, if it is made by cutting the
tter technique ("roulage sur natte"), in fact,
gives very different soft and shading impressions, with regular schemes but
very r pressure on the paste (cf. the pot published by Hugot 1963: pl. ilso Hurley
1979). In the concentric pairs of circles at the bottom ,ase, the
progressive reduction of the inner circle is obtained he compass pivot twice on
the same dot whenever the two points of fall along the same vertical line (fig.
9.4,7).
technique, recognized in some cases by Camps-Fabrer
(1966:467) and ed by Savary (1967:863), was never taken into consideration for
contexts. It was, on the contrary, well known in the Sudan, first 3 by Arkell
in 1949, but not recognized thereafter.
Technic~ue
s is, in a general sense, the same technique, but without
spacing s of lines, i.e., for each pair of lines the compass goes back to . the
dots of the last line impressed ("return" technique), ensuring :t
equidistance between the dots on the whole dotted surface, g a well balanced
smocking effect (fig. 9.4,4-7). In many cases the is so perfectly performed
that there are no clear technological ons. However, as already said, the
regularity itself was for us a iication. In addition, in most cases little
imperfections were which led to the discovery of this technique: the dots are
rly shaped, deformed by repeatedly going over them, often with t shaped points
(figs. 9.5,2;6,2); where the compasses did not go ctly the same point,
additional dots are present, intermediate and is to the pattern (fig. 9,5,1)
more frequently, inside the hole, the seen to be raised by the second
impression, which partially filled >revious one. If the rim, i.e. the beginning of the pattern is , a first line of small regular dots can
be noticed which is it from all the others, because it is the result of a
single :)n (fig. 9.6,1) (Barich, Mori 1970: fig. 12,2,6 11),
technique, unknown in the middle Nile Valley, is the most
id for decoration in the Saharan Neolithic. Similar, in principle, )cker
"guilloche", it was nonetheless never taken into consideration nt for
the impressed punctations, which were always considered to n made with normal
impression of combs or single points.
last two techiques are apparently behind all the
decorations known tations". They are
characterized by deep impressions, sometimes also on the back of the potsherds.
'_te the simple character of the foregoing technological
picture, the lsive decoration group is still rich and complex An internal ion,
classifying the different patterns and tools, therefore, would L, but it is
difficult to achieve.
frontal or lateral edge of a spatula or stick. which has
been found so far either of bone, distinction between the two techniques,
which, as on other elements, has appeared in some cases significance.
This is the kind of comb stone or potsherd. The already
said, also relies to be of very important
The so-called "basket" motif appears also to be
obtained with this technique (or with the following), not with the impression
of vegetal fibres - the latter technique ("roulage sur natte"), in
fact, gives very different effects: soft and shading impressions, with regular
schemes but very irregular pressure on the paste (cf. the pot published by
lIugot 1963: pl. XX; cf. also Hurley 1979). In the concentric pairs of circles
at the bottom of the vase, the progressive reduction of the inner circle is
obtained making the compass pivot twice on the same dot whenever the two points
of the tool fall along the same vertical line (fig. 9.4,7).
This technique, recognized in some cases by Camps-Fabrer
(1966:467) and suspected by Savary (1967:863), was never taken into
consideration for Saharan contexts. It was, on the contrary, well known in the
Sudan, first described by Arkell in 1949, but not recognized thereafter.
"Return" Technique
This is, in a general sense, the same technique, but
without spacing the pairs of lines, i.e., for each pair of lines the compass
goes back to pivot on the dots of the last line impressed ("return"
technique), ensuring
a perfect equidistance between the dots on the whole
dotted surface, providing a well balanced smocking effect (fig. 9.4,4-7). In
many cases the pattern is so perfectly performed that there are no clear
technological indications. However, as already said, the regularity itself was
for us a good indication. In addition, in most cases little imperfections were
noticed which led to the discovery of this technique: the dots are irregularly
shaped, deformed by repeatedly going over them, often with different shaped
points (figs. 9.5,2;6,2); where the compasses did not go over exactly the same
point, additional dots are present, intermediate and extraneous to the pattern
(fig. 9.5,1) more frequently, inside the hole, the clay is seen to be raised by
the second impression, which partially filled in the previous one. If the rim,
i.e. the beginning of the pattern is present, a first line of small regular
dots can be noticed
different from all the others, because it is the result of
impression (fig. 9.6,1) (Barich, Mori 1970: fig. 12,2,6 1).
which is a single
This technique, unknown in the middle
the most
The last two techiques are apparently behind all the
decorations known as "punctations". They are characterized by deep
evident also on the back of the potsherds.
impressions, sometimes
Despite the simple character of the foregoing
technological picture, the comprehensive decoration group is still rich and
complex. An internal subdivision, classifying the different patterns and tools,
therefore, would be useful, but it is difficult to achieve.
It must be said, in general, that within the four groups
the motifs result from the use of different tools, whose nature and aspect will
never be completely described and listed (suffice it to think of the endless
possibilities provided by vegetal or animal objects). In addition, the
inclination, the intensity and regularity of pressure applied to the tool and
the spacing of the impressions could produce such different effects in the same
kinds of decoration that the relationship between them is completely hidden, as
already noticed for instance by Camps-Fabrer (1966a:147). These features are,
therefore, the result of such accidental movements that they appear rather as
shades, often mixed on the same vase, fading one into the other and, in any
case, characterize almost every fragment. Hence, they will not be considered in
this chapter, the aim of which is to formulate a new perspective of ceramic
classification. For a detailed analysis of patterns the reader is referred to
the complete description of the Uan Muhuggiag materials by B.E. Barich in
chapter 7, this Volume.
4. THE TECHNIQUES OF DECORATION AT UAN
MUHUGGIAG: FREQUENCY DISTRIBUTION BY LEVELS
The table (Table 9.1) showing the frequency of the
techniques of decoration contains the four groups, a separate heading of
"varia" with rare cases of rocker combs, and a category of
unclassifiable sherds, which includes fragments decorated with alternate
punctations in which the "return" technique was not clearly
understandable (4). Therefore, when clustering the whole group of punctations,
in contrast with the group of plain rocker, unclassifiable sherds are included
in the first group.
It is to be noted that, as far as the plain rocker is
concerned, the presence of this decoration on the rim has been considered as an
additional occurence.
A detailed analysis of the occurrences is difficult and
not fully reliable because of the scarcity of the materials, except in one case
(less than 50 units per level). Any conclusion has, thus, a merely indicative
value.
Sector B provided more
material and a greater stratigraphic extension. Both Sectors, however, show the
same trend, especially in their lower levels. A
gradual increase of the rate of the rocker technique can be seen throughout the
stratigraphic development (levels 2d, 2c, 2b, 2a in B;
(4) In most cases, from the extreme regularity
of the pattern, we must infer that the "return" technique was
applied. The simple alternately
pivoting tool, in fact, gives less regular patterns.
levels 2c, 2b, 2a in A) maintaining, in both Sectors,
almost the same ratio between the two main groups.
In the highest levels, however, there is an inconsistency
between level 2 of Sector A and the corresponding 2 in Sector B. The latter is
probably to be seen as part of the underlying level (i.e. 2a), since the
inversion of the trend, with an exceptionally high rate of alternate
punctations, characterizes level 2 in sector A and level la in sector B. This
is probably just a matter of labelling, not the result of a different
situation. Actually, the two
Sectors compare very well, which confirms the reliability of the analysis.
Generally speaking, it can be said that the
"return" technique is the most common in the site throughout the
chronological development. It also prevails on the simple alternate stamp. The
increase of the plain rocker technique accordingly appears to follow a phase of
"return" technique ("fish scale"). In other words, the
diffusion of this technique could be linked to its possibility of providing
kinds of smocking patterns like the "return" technique. The same situation seems to be attested at
Wadi Athal (Barich, Mori 1970); at Ti-n-Torha a similar situation is evidenced
in the lack of rocker technique in the lower two levels, while this decoration
accounts for 13-18% in the latest contexts (Barich 1974: tab. X).
COMPARISONS AND CONCLUSIONS
As far as we can gather from lists and descriptions
relying on a different typological system, the foregoing techniques of
decoration are typical of the Neolithic Sahara, where almost in every region
they chronologically follow contexts characterized by the dotted wavy line ware
(sensu Arkell; cf r. note 5) and by the rocker comb decoration.
As far as the Saharan occurrences are concerned, however,
we refer to the comparisons made by B.E. Barich on the basis of the complete
contexts, limiting the analysis here to the most eastern, non Saharan,
evidence.
As far as the Nile
Valley is concerned, the alternate impression of twotoothed instruments occurs
already in Mesolithic contexts, associated with wavy line, dotted wavy line and
rocker comb decoration (Caneva 1983). Its
rate appears to increase during the early Neolithic (Shaheinab culture)
becoming then the exclusive impressed decoration during the late Neolithic (cf.
Kadada: Geus, Reinold 1979).
The shift from the rocker comb to the pseudo-rocker double
point is well evidenced in the pottery from central
stamp) (fig. 9.1) (5).
The same applied (as far as the photographs show) in the
Ennedi, where a kind of dotted wavy line with small, sharp waves, associated
with rough triangular punctations, is found to overlap the rocker decoration
and the classical dotted wavy line (levels B and A at Soro Kezenanga; Bailloud
1969: 38, fig. 3, A) (6).
In the middle Nile Valley the plain rocker follows the
same trend: almost lacking in the Early Khartoum contexts (only one fragment is
published in Early Khartoum, Arkell 1949: fig. 79, 3; not one was found at
Saggai), it becomes a more and more common technique of decoration during the
Neolithic phases. Often used as rim decoration, it is confined to the rim of
otherwise undecorated vases in the latest cultures (Arkell 1953: fig. 31,
40-42).
However, other elements differ considerably in the two
areas; not only paste composition and vessel shapes (as far as we can infer
from such small fragments, rims are more varied in the Sahara) but also the
decorative motifs, which are quite diversified and usually well spaced in the
Sudan. Here, in addition, the designs are less deeply impressed into the paste,
never visible on the inner surface. But the most impressive difference is the
role that these techniques hold in the two areas: exclusive in the Sahara,
medium in the Sudan, where they are always associated with other kinds of
surface treatment, such as incision, brushing, red slipping, combing-and-washing
(ripple ware), which give the ceramic production a well varied character in
comparison with the stereotyped aspect of the corresponding Saharan production.
The "return" technique is almost unknown in the
Nile Valley (except for some cases of "fish scale" motifs).
Therefore, the particular effect of oblique alignments and lozenge spaces,
which is so characteristic of the Sahara, is there completely unusual.
From this comparison there is, therefore, a general
impression of parallel technological development through similar stages, which
were already partially recognized by Arkell for the Khartoum province (1949)
and by Bailloud for the Ennedi (1969). The
sequence of these stages now appears to extend all over the Sahara: from wavy
line to dotted wavy line, to rocker comb impressions, to alternate dots
together with plain rocker.
These stages not only are more or less contemporary in
both areas, but also are associated with broadly similar cultural contexts,
i.e.
The misconception concerning the dotted wavy line ware was
already known by Arkell, who tried then to emphasize (apparently too late) the
difference between the two kinds of this decoration recognized in the Sudanese
prehistory (1972).
These are the levels to be compared with Sudanese
Neolithic cultures instead of the lower ones, where both archaeological
materials and C14 datations perfectly correlate with the Early Khartoum
cultures.
"Mesolithic" economy and tool kit, with
documented hunting/fishing and gathering activities as far as the more ancient
series (wavy line, dotted w.l, and rocker comb) is concerned,
"Neolithic", pastoral economy, with reduced lithic tool-kit (both in
number and variety, cf. Barich 1974) associated with the pottery having
alternate dots.
The dates of the first group range between 6000 and 4500
B.C., both in the Sahara and in the Nile Basin (cf. Shabona, Sorourab II,
Saggai), while the more recent group is dated 4000 in the Acacus, 3000-3500 in
the Ennedi, 2500-2900 B.C. in the Nile Valley (Kadada). The Saharan dates appear to be somewhat earlier than the
Sudanese, but this is due to the lack in the Sahara of something really
comparable to the Shaheinab culture which is dated 4000-3500 B.C. In any case, the series is very incomplete
for such an extensive area, and no reliable hypothesis can be developed on the
origin of one or the other cultural complex (7). Moreover, a valid comparison should take into consideration
the whole complex of elements forming a cultural context. The decoration of pottery is merely part
of one of these elements. It has
been considered an important one, often the only one available for study of the
Saharan prehistory (Camps 1974: 235), but its content of information is only
indicative of the relationship between the two areas considered.
This concept, always at
the basis of Barich's work in the Acacus, was not widespread among Saharan
prehistorians, as Maitre has many times stressed. He pointed out the extremely varied internal faceting of
the Saharan Neolithic (he says "Neolithics") and the poor possibility
of comparison between the two areas (1972). However, his perspective was
biased: behind his obstinate defence of the Saharan autonomy, one can guess
that there may be a doubt as to the autonomous cultural development of one or
the other areas, and that the question is which one influenced the other.
Moreover, he is involved in an incorrect analysis of the Sudanese contexts,
mistaking Shaheinab for Early Khartoum. Hence
his evidence, which is based on the chronological priority of Delebo (5250
B.C.), becomes a refutation once Delebo is compared with its right homologue,
that is Saggai (5500 B.C.).
Actually, the close parallelism and the contemporaneity of
the stages, as well as the local diversities, do not suggest any particular
priority. On the contrary, they emphasize the possibility of long,
uninterrupted reciprocal relationships between the two areas.
These relationships, as seen through the pottery, seem to
have always existed, lasting until very recent times (cf. the identity between
the pottery of the Nubian C group and that of the Iron Age in the Borkou;
Courtin 1969). We assume,
therefore, that such relationships must be considered in a normal network of
contacts among fairly mobile groups inhabiting the same environment and
practising the same activities, rather than as episodic events requiring a
specific explanation,
Camps is rather inclined towards a Saharan origin of the
African "neolithisation" (1974: 224), but he bases this assumption on
the comparison between the farming communities of the Fayum, which, of course,
have a very recent date (4500 B.C.), and the Mesolithic sites of the Acacus,
about 2000 years older.
The different local features, especially the different
associations, are easily explained as autonomous developments and utilizations
of broadly common schemes. The
data to quantify its significance in the internal development of a single
context are still lacking. In
addition, the actual contexts are very few, since the findings are either
sporadic or old in most cases. This
situation, in fact, could account for the lack of actual syntheses of both the
Sahara and the Sudan.
Dipartimento di Scienze Storiche, Archeologiche e
Antropologiche del 1'Antichita, University of Rome "La Sapienza", Via
Palestro 63, Rome, Italy.
REFERENCES
ADDISON F. 1949, Jebel Moya, Oxford University Press: London.
ARAMBOURG C., BALOUT L. 1952, 'L'ancien lac de Tihodaine et ses gisements prehistoriques', Actes du II eme Congr. Panafr. de Prehist., Alger: 281-292.
ARKELL A.J. 1949, Early Khartoum, Oxford University Press: Oxford.
ARKELL A.J. 1972, 'Dotted Wavy-line Pottery in African Prehistory', Antiquity, 46/183: 221-222.
BAILLOUD G. 1969, 'L'evolution des styles ceramiques en Ennedi (Republique du Tchad)', Actes du Premier Colloque Intern. d'Archeol. Africaine, Fort Lamy: 31-45.
BARICH B.E. 1974, 'La serie stratigrafica dell'Uadi Ti-n-Torha (Acacus, Libia). Per una interpretazione delle facies a ceramica saharo sudanesi', Origini VIII: 7-184.
BARICH B.E., MORI F. 1970, 'Missione Paletnologica Italiana nel Sahara Libico. Risultati della Campagna 1969', Origini, IV: 79-142.
CAMPS G. 1968, 'Amekni: neolithique ancien du Hoggar', C.R:A.P.E., 10 (A.M.G.: Paris), 230 p.
CAMPS G. 1974, Les civilisations prehistoriques de 1'Afrique du Nord et du Sahara, Doin: Paris.
CAMPS-FABRER H.1966, 'Matiere et Art mobilier dans la Prehistoire nord africaine et saharienne', Memoire du C.R.A.P.E., 5 (A.M.G.: Paris), 574.
.
CAMPS-FABRER H.1966a, 'Sur quelques techniques decoratives de la ceramique impressionnee saharienne', Congres Prehistorique de France, Compte Rendu de la XVIIIeme Session, Ajaccio: 143-154.
CANEVA I. 1970, 'I crescenti litici del Fayum', Origini, IV: 161-204.
CANEVA I. ed. 1983, 'Pottery Using Gatherers and Hunters at Saggai (Sudan): Pre-conditions for food-production', Origini, XII: 7-278.
CANEVA I. 1984, 'Typological Notes: the Sudanese Case', Paper presented at the II Intern. Symp. on the Late Prehistory of the Nile Basin and the Sahara, Poznan.
COURTIN J. 1969, 'Le neolitique du Borkou, Nord-Tchad', Actes du Premier Colloque Intern. d'Archeol. Africaine, 1966: 147-159.
GABRIEL B. 1981, 'Die Ostliche Zentral Sahara in Holozan - Klima, Landschaft und Kulturen (mit besonderer Berucksichtigung der Neolitischen Keramik)', in Prehistoire Africaine, Melanges offertes au Doyen Lionel Balout, reunis par Roubet C., Hugot H.J. et Souville G., A.D.P.F.: Paris, pp. 195-211.
GEUS F and_ REINOLD J. 1979, 'Fouil le de sauvetage a el Kadada (Soudan): la campagne d'avril 1976', C.R.I.P.E.L., 5: 7-159.
HAALAND R. 1981, Migratory Herdsmen and Cultivating Women, Mim: Bergen.
HAYS T.R. 1974, 'An Element of Nilotic Diffusion', South Afric. Archaeol. Bull., 29: 27-32.
HUGOT H.-J. 1963, 'Recherches Prehistoriques dans 1'Ahaggar nord-occidental', Memoire du C.R.A.P.E., 1 (A.M.G.: Paris), 206 p.
HURLEY W.M. 1979, Prehistoric cordage Identification of impressions on pottery', Manuals on Archaeology, 3, Aldine: Washington.
MAITRE J.P.1972, 'Notes sur deux conceptions traditionnel les du Neolithique saharien', Libyca, XX: 125-136.
MOHAMMED ALI A. 1982, The Neolithic Period in the Sudan c.a. 600025000 B.C., Monographs in African Archaeology No. 6: Cambridge.
ROSET J.P. 1978, 'Poteries neolithiques du Tenere. I, la region de 1`Adrar Chiriet', Cahiers O.R.S.T.O.M., ser. Sc. Hum., XV, 4: 379-406.
ROSET J.P. 1983 , 'Nouvelles donnees sur le probleme de la neolithisation du Sahara meridional: Air et Tenere au Niger', Cahiers O.R.S.T.O.M., ser. geol., XIII, 2: 119-142.
SAVARY J.P. 1967, 'Elements remarquables du Neolithique saharien', Bull. de la Soc. Preh. Frang., LXIV: 859-864.
Back to Library | Contents | home