AR62. 11 November 2006.
Copyright © 2006 by Kevin Sharpe and Leslie Van Gelder. All rights reserved.
In process. Detailed report on the 3/4?? panels. Zipf #3.

 

Applying Zipf, Internal, and Forensic Analyses to Three Panels of Finger Flutings in Rouffignac Cave, France

Applying Internal, Forensic, and Zipf Analyses to Three/Four++??? Panels of Finger Flutings in Rouffignac Cave, France // Establishing a Notational Intention behind Two Panels of Finger Flutings in Rouffignac Cave, France // Evidence for Notation in the Upper Paleolithic

 

 

by

Kevin Sharpe

Graduate College, Union Institute & University, Cincinnati, Ohio, USA
Harris Manchester College, Oxford University, Oxford, UK

10 Shirelake Close, Oxford OX1 1SN, United Kingdom
ksharpe@ksharpe.com
www.ksharpe.com

 

and

Leslie Van Gelder

College of Education, Walden University, Minneapolis, Minnesota, USA

 10 Shirelake Close, Oxford OX1 1SN, United Kingdom
leslievg@ksharpe.com


ABSTRACT.

This paper uses an empirical methodology to examine two clusters of a variant of the Rugolean Form of finger fluting found in Rouffignac Cave, France, asking how they were made and what they might reasonably tell about the people who made them. An initial result of this approach is to indicate the ages, heights, genders, and number of fluters involved. Further, the application of Zipf’s Law from communications theory, based on the number of fingers used in each unit of a cluster, suggests that these flutings were a form of inter-subject communication. Possible conclusions from these data and inferences are explored.

KEY WORDS.

??? Finger flutings, methodology, parietal art, prehistoric art.

CONTENTS.

Introduction. 6

Mirian. 7

Kirian. 7

Rugolean. 8

Three Methods. 8

Internal Analysis. 8

Forensic Analysis. 8

Zipf Analysis. 9

Results. 9

Chamber …, Panel 1. 9

Internal Analysis. 10

Forensic Analysis. 10

Zipf Analysis. 11

Chamber …, Panel 2. 11

Conclusion about this panel: 11

Internal Analysis. 11

Forensic Analysis. 11

Zipf Analysis. 12

Chamber …, Panel 1. 12

Internal Analysis. 12

Forensic Analysis. 13

Zipf Analysis. 13

Chamber …, Panel 2. 13

Internal Analysis. 13

Forensic Analysis. 13

Zipf Analysis. 13

Chamber …, Panel 3. 13

Internal Analysis. 13

Forensic Analysis. 13

Zipf Analysis. 14

Discussion. 14

Conclusions and Questions. 14

Acknowledgements. 14

References. 15

 


 

·         It’s striking that there are no undulations in these two panels (Mammoths of Discovery and the Patriarch). Re other indicators of different intentions.

·         What isn’t there in the cave and what isn’t there in each panel. E.g., lines circling rocks in A1 and E, but nowhere else we’ve studied (½ circling, twice, on a wall). Zigzags (undulations) not in some panels. Re other indicators of different intentions.

·         Continued validity of the argument for Rugolean’s being writing with children not making lines in Chamber E or in Breuil. Relates to forensics.

·         Another measure, called Shannon entropy, may help gauge the complexity of the communication, if indeed it be that. To try it out.

But another aspect of information theory can take you further than just knowing a species has the potential for advanced communications. Without understanding what is being said, you can find out just how complex its communications are.

Claude Shannon, the father of information theory, pioneered the approach. Shannon employed the term ‘entropy’ to size up the complexity of a communication system. It is not just a concept borrowed from thermodynamics: Shannon introduced a whole hierarchy of entropy levels (see Graph). Zero-order entropy measures the diversity of the communicative repertoire – how many different types of elements (letters, words, phonemes, whistles, barks and so forth) make up the signal. In written English, for example, everything can be represented by 27 characters – 26 letters and a space. The entropy value is the logarithmic value of the number of elements. The researchers used base 2 logarithms because the binary digit or bit, is the most familiar form for information (though any base will do), making the zero-order entropy of English to be 4.755. Next in Shannon’s scheme is first-order entropy, which measures – again, using logarithms – the frequency of occurrence of each element within the language.

The higher entropy levels, second order and up, relate to the notion of ‘conditional probabilities’: once you have seen a particular sequence of elements, what are your chances of predicting the next element in the series? If, for instance, you know the first and second words of a phrase, the third-order entropy tells you (in logarithmic form) the odds of guessing the third word correctly. Analyses of English and Russian suggest that these languages show evidence of 8th or 9th-order Shannon entropy, meaning that when presented with a string of eight words, you have some ability (slim but non-zero) to predict what the ninth word might be. After that, though, all bets are off. If you want to guess what the 10th word is, the previous nine are of no value.

Primitive communications, such as chemical signaling systems employed by cotton plants, don’t go beyond first-order Shannon entropy. That means there is no discernible connection between signals – knowing one doesn’t help you predict the next. Adult squirrel monkeys, on the other hand, show second or third-order Shannon entropy: their ‘language’ has some predictability in its structure, but not much. So far, the research of McCowan, Hanser, and Doyle indicates that dolphin whistles bear signs of 3rd or 4th-order Shannon entropy.

This kind of analysis can reveal differences in communication complexity between the species. Interpretation of such graphs is complicated, but it does show that human languages have broadly similar communication potential, and that other animals have different capabilities, the limits of which have not yet been fully explored (Nadis 2003: 38).

·         Do forensic analyses (perhaps there will be other important data here for the next study).

·         Shannon’s entropy.

·         Do I need a forensic analysis? Maybe not at this stage – for example different fluters – could go into the questions for the future.// to do a full forensic analysis in this paper may take too much space and documentation. Do I need it at first?

·         Perhaps the curved ||  || is a form or element of the grammar (in MoD Panel).

The following are observations made in the cave in March 2007. Whether they are relevant in this paper or the fuller version or something else (e.g., when doing forensic analyses of these 2 panels) needs to be decided.

Mammoths of Discovery:

·         Both mammoths appear to be fluted by the same 16 mm fingered person who also appears to have done all of the flutings surrounding the mammoths (i.e., th doubles and quadruples). The tops are confusing, are not 41. In one instance, they appear male. Often a curvature to the tops so hard to say.

·         Need to experiment with children (especially and adults) to see the vertical proximity of the pinky to the other fingers (would be on young kids versus adults). The ‘pink distance’ may be another good indicator of age and individuality.

·         50 (48) just after the ‘X’ gives 2=29 and 1=14. Above it a fingered unit of 3=38, then another of 46 that gives 1=16-17.

·         After looking, 38 and 50 both seem female.

·         The panel with curved ends has in both cases 2-fingered units on both sides. Units in between done by different people, which implies not the same person.

·         Perhaps the curved ||  || is a form or element of the grammar.

·         ‘Profiles matter.’ Judge by widths and then if same width, see if there are other distinguishing characteristics.

·         Rugolean units of RH end of panel are 41 (i.e., 41 is present right after the bear scratches). Tops are 41 are obvious.

AB Rugolean (i.e., Patriarch) Panel:

·         RH end mammoth. Inside is a 3=44 unit with very rounded tops (1=14). Back of mammoth is 1=15; 2=34S.

·         In the panel: 34, 41, 38 by widths and profiles. 38 appears to have much thinner fingers than 41 (perhaps).

·         RH mammoth: 3=36 and 34. Single finger on back of mammoth: 14 and 13. Index on front? is 13. 1 of 3.

·         LH mammoth: 3=43, 2=31, 1=15 (back). If 44, did 41 do the mammoths at both ends of the panel?

·         Mess to left of mammoth has 34 in the middle and 28 as well.

 

·         Maybe do the 3rd paper before the 2nd – or at least refer to it. I need to spend some time working out the relevant content of each.

·         Use of Shannon entropy measure.

·         Forensic analysis and its influence (or should this be the 3rd paper?).

·         Chamber E for 3rd paper.

·         Do yet another Zipf paper for Gargas – 2 panels plus baton.

·         Forensics may explain why the Patriarch Panel does not produce a gradient of -1.

·         Fuller discussion of Zipf.

·         Fuller discussion on notation, etc.

·         I think this paper still needs to focus on the 2 panels in the 1st paper.

·         Many of the future directions will come out of the 1st paper.

·         So the 3rd paper is a fuller analysis of 2/3/4 panels in Rouffignac Cave. Maybe the forensics go in here.

Introduction

A full analysis of the four+ panels in Rouffignac Cave, building on AR52 and applying AR53. To see what results from dialogue between the published Zipf analysis of Panels … and the internal and forensic analyses in previous paper AR52.

·         Use Zipf’s Law on Blombos ochre.

·         Do the number of units repeat between clusters? If so, does this give a similarity to Kirian flutings?

·         Have I written more extensively about the Rugolean form elsewhere? Seems to be very little here. Perhaps in the forms paper. Also there must be more notes on them other than those from November 2003.

·         Add stuff from previous papers, e.g., on the forms in Rouffignac.

·         Looking for meaning-repetition-structures across the panels: use Zipf’s law to find them.

·         We see a commonality of 6 units of 4 near the LHS of 2 Rugolean clusters near the Mammoths of Discovery and the Patriarch.

·         It’s striking that there are no undulations in these two panels (Mammoths of Discovery and the Patriarch).

·         Continued validity of the argument for Rugolean’s being writing with children not making lines in Chamber E or in Breuil.

·         What isn’t there in the cave and what isn’t there in each panel. E.g., lines circling rocks in A1 and E, but nowhere else we’ve studied (½ circling, twice, on a wall). Zigzags (undulations) not in some panels.

Introduction including to the cave, our work (history of publications).

Historical, geographic, cave layout

Update from AR50 on Kirian Form

Prehistoric finger flutings (the lines that human fingers leave when drawn over a soft surface) occur in caves through southern Australia, New Guinea, and southwestern Europe, and were presumably made over a considerable time span including some or all of the Upper Paleolithic. Most are not obvious figures or symbols; flutings of this type are termed ‘severines.’ Rather than speculate as to meaning or the development of types of severines over time, this paper continues to establish an objective and experimental approach to the lines (Sharpe and Lacombe 1999; Sharpe, Lacombe, and Fawbert 1998; 2002; Sharpe and Van Gelder In Prep.).

The following terminology may help when discussing flutings:  ??? say draws on … reference for terms relevant to this report and the development of new terminology. ???

·         As said above, the word fluting refers to a line drawn with a finger.

·         The phrase graphical unit (or, abbreviated, the word unit) refers to flutings drawn with one sweep of one hand or with one finger (Marshack 1977). (Previously, this was referred to as a ‘stream’ (for instance, in Sharpe, Lacombe, and Fawbert 2002); the terminology has been changed to avoid possible interpretative inferences.)

·         The word cluster labels an isolatable group of units that exhibit a unity, for instance because they overlay each other.

·         The word panel of clusters …. ???

The term fluting applies to line markings made with fingers, and the term engraving refers to line markings made with a tool. Within engravings, a difference exist between scratches (animal claw marks), incisions (lines that humans make with flint or other piece of rock), scorings (lines that humans make with a stick), and bone marks (lines that humans make with a bone).

The term severine is suggested for line markings that do not participate in the figurative part of a definitive figure or demonstrable symbol or sign, equivalent to Marshack’s term ‘meander’ (Marshack  ??? ) but without the restrictive overtones of this word. Thus, the category ‘line markings’ not only comprises flutings and engravings but, coextensively, also severines, figures, and symbols.

The following three forms of fluting have been introduced and possibly established for finger flutings found in Rouffignac Cave, France (Sharpe and Van Gelder Preprint):

Mirian

One thing defines a cluster of finger flutings as of the Mirian Form: lower-body movement on the part of the fluter. ‘Lower-body movement’ means that the people who fluted the walls or ceilings in the Mirian Form not only sometimes walked or otherwise moved their legs while fluting (thus the lines may extend beyond the arm range of a stationary fluter), but almost always moved their bodies from their hips to create the flutings (as opposed to only moving their upper bodies).

Kirian

Two things define a cluster of finger flutings as of the Kirian Form: the fluter standing in one spot while fluting a unit, and each unit comprising only one line. The people who fluted the walls or ceilings in the Kirian Form stood still and marked with one finger.

Rugolean

Two things define a cluster of finger flutings as of the Rugolean Form: the fluter standing in one spot while fluting a unit, and most units comprising more than one line. The people who fluted the walls or ceilings stood still, moved their upper bodies, and marked mostly with more than one finger at a time. The fluter may have moved between making units, but stood stationary for each unit.

This paper continues the stylistic analysis of flutings of the Rugolean Form, particularly the flutings found in Gallery … and … (otherwise known as the … respectively) of Rouffignac Cave.

Previous work on Rugolean Form in Rouffignac and elsewhere

·         Look for units touching or overlaying to see the direction for making the units.

Give the impression of multiple markers working in concert.

Three Methods

Overall questions.

Specific questions. Introduction to the three methods. // Overall methodology, specific methodologies, comparison with other methodologies

Internal Analysis

 

Forensic Analysis

Specific questions for this form

·         Do adult hands make long horizontals along the walls, or only children?

·         Seems to be RH 3s and LH 4s. correct?

·         In groupings, are there always RH 3s and LH 4s?

·         Number of fluters.

·         Can we distinguish individuals among Rugolean sets?

·         Fingers splay when stretching. The further out from the center, the more the fingers splay.

·         Need notation as to where the marker stood.

1.      Need where people stand, and relative positions of them, which implies

1.      White space,

2.      Continuous movement of marker from place to place, or

3.      Different acts.

2.      Mark on photo of area where a cluster made from.

3.      There is a sense of the relative placement of clusters.

·         Different widths imply different persons?

·         Look at the Rugolean lines specifically for gender.

Overall (Update this):

To understand the fluting creation process and to know about the behavior and biology of the fluter(s).

·         Distinguishing between individual fluters

·         How many fingers

·         Left or right hand

·         Gender of fluter

·         Height of fluter

·         Age of fluter

·         Direction each fluting made

·         Direction in which the flutings were applied

·         Order in which the flutings were made

·         Layout of cluster and panel of clusters

·         Height of cluster above original floor

·         Addition of things (such as clay and charcoal) over the fluting

·         What movements of the fluter were involved in the marking process

·         Use of other marking techniques besides finger fluting

·         Condition of fluting: comparing old and new ones

 

·         Generally upright. Markings show evidence of stationary line marker. Length of lines tends to be determined by the arc of an arm moving in one gesture (generally from top to bottom or from the side across to bottom).

·         Leslie’s trying with finger paints when away from the center, even the starts of units show a curvature indicative of the hand L or R.

Zipf Analysis

Results

·         Research results

Chamber …, Panel 1

To the right of the Patriarch. Assume is Breuil.

·         Are the top (narrow) lines Jasperian versus Rugolean? But kids doing their thing while adults were doing something else?

·         We see a commonality of 6 units of 4 near the LHS of 2 Rugolean clusters near the Mammoths of Discovery and the Patriarch.

·         It’s striking that there are no undulations in these two panels (Mammoths of Discovery and the Patriarch).

Internal Analysis

·         Around the Patriarch:

o       For Rugolean lines: put the over/underlay analysis further back. Isolating the number of individuals may be more important.

Forensic Analysis

·         Around the Patriarch:

o       One person with R and L hands and space between.

·         Are all the units here (the Patriarch lines) right handed? If they are, why the curve to the left? Experiment with this.

The following notes are from the Abbé Breuil chamber, but whether they pertain to the first or second panel is unclear. I also recorded the measurements as cm but I guess they’re meant to be mm.

Flutings – end – LHS

1.      40 wide at top 3 fingers – possible RH
L – 2 mm C – 5 mm – R relative heights

2.      to right of (1)
33 wide – tops obscured

3.      to left of (2)
36 wide – LH possibly – tops obscured

4.      to left of (3)
36 wide – top obscured

(1) – (4): All measurements unclear
(5)+: inaccurate by about 1-2 mm

5.      RH 38 mm wide
3:5 relative heights

6.      ~3 units to left of (5)
?RH 38-39 mm wide
2:5 or 6 relative heights

7.      2 cross units both 30 mm under (5) and (6)
obscure beginnings
Much harder to see the units clearly enough to measure it than visually to distinguish it

8.      to right of (1), also 40 mm wide
top obscure

9.      to right of (8), 22 mm is 2 digit unit
top obscure

10.  to right of (9), 22 mm is 2 digit unit
top obscure

11.  3 digit at 38 mm
top obscure

12.  Kirian – tops obscured by bear scratches

13.  horizontal unit 35-36 mm (3 central fingers), 45 mm (4 fingers)
top obscure

14.  2 fingered, 20 mm vertical, to right of (12) and over (13)

15.  2 fingered, 25 mm vertical, to right of (12) and over (13)

16.  horizontal – 26 mm – 3 fingers
beginnings obscured
ends further to right of (15) over a set of Kirian lines looking like thumb prints – deep – fan out – individually done.

·         Appear to be a lot of hands here.

·         Very difficult to tell the relative heights: the act of measuring is very difficult.

Zipf Analysis

Chamber …, Panel 2

To the right of the Mammoths of Discovery

·         We see a commonality of 6 units of 4 near the LHS of 2 Rugolean clusters near the Mammoths of Discovery and the Patriarch.

·         It’s striking that there are no undulations in these two panels (Mammoths of Discovery and the Patriarch).

Conclusion about this panel:

·         Has some single lines following bear scratches as in other Rugolean panels.

Internal Analysis

·         First mammoth panel: x’s: LH unit under RH unit.

·         The pairs in the mammoths of discovery:

o       Tusks cut over everything.

o       The very left lines aren’t paired.

o       To the right of the X cluster there are smaller clusters; a V, etc.

·         The underlying cluster/unit is a large horizontal X.

Forensic Analysis

·         Around the Patriarch:

1.      Looked at finger widths => more than one individuals (between 3 and 5).

·         First mammoth panel: tusks over the paired lines => mammoth over the pairs and quadruple units.

·         At the X itself, the side starting on the left is done with the left hand, the side starting on the right is done with the left hand and is over top of the left side (see the thumb marks as well).

·         The second right unit has a different curvature and looks less weathered.

·         Looking to measure finger top height:

1.      Can’t do a unit that starts curved.

2.      The LHS doesn’t give a good finger top height.

3.      The RHS person didn’t have big relative height differences.

4.      Better to go with where the person stood.

5.      Or with finger widths.

6.      Problem also with the wall curvature where the lines start.

7.      Even a photo wouldn’t help measure the relative heights as the information isn’t there.

8.      Stretching would also mess up the relative heights.

·         They have better finger tops than the others.

·         The quads curve differently from the pairs.

·         The pairs in the mammoths of discovery:

1.      Very small lines, much smaller than Leslie’s.

2.      The very left (long?) seems splayed.

3.      For the pair units, there is one spot for making them.

4.      For the quad units, there is also one spot (the epicenter) for making them.

5.      There’s about 3 feet between the epicenter for the pairs and for the quads.

6.      The different line widths between the pairs and the quads say they were done by different people. They could be contemporaneous because they’re 3 feet apart.

7.      The epicenter at the X is much further away, about 5 feet from the epicenter of the quads to the epicenter at the X.

·         X panel: Séverine could draw most of it while standing at the X.

·         RHS has right tilt and LHS has left tilt. Curvature => where standing.

·         Did more than one person make the lines at the two mammoths?

·         First mammoth panel: Looks like 3 clusters in the panel, especially the middle one with the small rocks on either side and two x’s in the middle – left-hand units have LH leaning angle and RH units have RH leaning.

·         By the 2 mammoths for finger heights.

·         Did more than one person make the lines at the 2 mammoths?

Zipf Analysis

Chamber …, Panel 1

In the Via Sacra, right hand wall???

·         Via Sacra:

o       Are some Rugolean lines and one large concentration of them.

o       Rugolean can include horizontal lines; just, not walking.

Internal Analysis

Forensic Analysis

·         Via Sacra Rugolean Panel:

o       28, 34-35, 42, 52 mm widths.

o       Many of the 28s are high up (a child being helpd aloft?).

o       Large ones are sometimes splayed.

o       Horizontal lines are of different widths.

Zipf Analysis

Chamber …, Panel 2

In the Via Sacra, left hand wall???

·         Via Sacra Rugolean II:

o       Horizontal lines are Rugolean.

Internal Analysis

Forensic Analysis

·         About 36 mm + others.

Zipf Analysis

Chamber …, Panel 3

Internal Analysis

Forensic Analysis

In the Via Sacra, moving left to right????, some widths:

1.      In ‘tectiform-type figure’:
40W – sets of 3 lines – consistently.

2.      To right of this, sets are of 40W

3.      Small units of 40W

4.      40W

5.      But beneath it is a 30
Lighter sets are 30

6.      40W

7.      30W to right

Seems as if older marks – fainter – are 30 – are higher marks including very high out-of-Leslie’s-reach marks.

Over these in bigger composition – are 40 – lower marks.

In this area, flint nodules missing (don’t appear to be elsewhere) with some evidence of forcible removal.

Zipf Analysis

Discussion

           Having done this with one cluster of lines of a particular style, seeing how another cluster of the same style stacks up against it.

           Can we break a cluster of structured units into ideational units (called motifs or something)? If so, how do the ideational units of one cluster compare with those of another cluster? (Maybe this is as close as we can get to meaning.)

           Then compare this from Rouffignac with the same analysis from another cave nearby and one from some distance away.

·         Are there repetitions of structures?

·         Including comparison of panels

Conclusions and Questions

Specific conclusions, general implications, further questions arising. Include other places in Rouffignac Cave that could do these three analyses on. Order these points by the questions originally asked.

·         We need a panel or cluster with enough clarity to look at. Which leads me to ask about ignoring certain lines while including others. Ignore underlying ones? NO. Ones with older weathering (fainter)? Different hands made them?

·         3’s and 4’s might not be significantly different if the 4 is a drag at the end or at corners. Note to make a 2 or a 1 or a 3 or a 4 with intention (i.e., holding fingers up) is more significant.

·         ??? Chamber A2, A sector 3. What did we find there?

·         Junction of A1 and A2, and into A2:

o       Some Rugolean lines are high up at the junction on the RH side (a person would have had to climb up on the flint nodules to have made the marks).

o       Some long horizontals along the RH side wall.

o       Space gets lower and lower the further down A, and yet there isn’t evidence of much on the ceiling (a lot of graffiti).

·         Measure widths of Rugoleans at the entrance to A2.

·         A1/A2 junction: 5 units fluted @ 48 mm for 3 fingers.

·         A cross over from flutings to engravings: Frederic Plassard -- 4 lines in the verticals of tectiforms he thinks is significant, in both engraved and fluted tectiforms. Is this a key?

·         The next big questions beyond the adequacy of this methodology and analysis concern translation.

o       That may not be possible.

o       Some leads: Ogham, etc.

o       May be very cultural, time, and geography dependent

o       The making of severines may be fairly universal however.

Acknowledgements

 

References