History Forum / General / Archaeology / January 2007

Ah, you weren't being clear. I didn't realise you meant here.

Doug

That there was, by definition, no such thing as vowelless Ogham was
one of the tenets of the arguments used against Larry Athy and one of
the foundations of the claim that led to your recent remark of Garbage
in - Garbage out.

Eric Stevens

That's correct but only to a point. You are describing the characters
of 'scholastic ogham' which is generally inscribed on a flat surface
and uses a stem line as a reference. As I understand it, this is the
later form of Ogham.

  ------

I've tried not to get to this stage but now I don't think I've got any
alternative. I don't claim to be knowledgable in Ogham so what follows
is based on my understanding of Larry Athy's argument. I can't argue
in favour of any other parts of it other than internally consistent
logic. Nevertheless I respect his views as the product of more than
twenty years work with the widest possible body of known Ogham
inscriptions from all parts of the world.

One of the problems with Larry's paper is that there is much by the
way of introductory and explanatory material that he left out. The
background to what he did, why and how he did it is something which
emerged only in the many thousands of articles in the arguments in
sci.archaeology which followed. I started to put it all together but
the threads were so complicated that I never finished.

Larry Athy described basically three forms of Ogham and described them
as 'proto-Ogham', 'Classic Ogham' and 'scholastic Ogham'. I know a
number of people (including Searles Odubhain and John Byrne) objected
to this but this terminology is not unique to Larry. It seems to have
started with Barry Fell and has been adopted/accepted by a number of
people in North America including David H. Kelley.

Although it appears to be claimed that there are some shifts in
phonetics as Ogham progressed from one form to another, the three
forms of Ogham have much in common.

Ogham has always been a system of writing employing a reference line
and short and long strokes. The reference line may be vertical or
horizontal. The short strokes start at the reference line and may
extend above the line or below the line. The long strokes cross the
line. There is also a group of five special characters which are
always carved in a fixed relationship to the line. There are apparent
conventions for the reading of the inscription according to whether
the refrence line is horizontal or vertical or circumscribing the
circumference of a slab of stone.

All descriptions of earlier Ogham are complicated by the fact that
most of the early inscriptions are on the corner edge of a stone as
shown in http://encycl-celt.ifrance.com/ogham.jpg with the result that
the two parts of the long lines were on two adjacent sides of the
stone. Consequently the two parts of a long stroke could never line up
and frequently were carved at different angles to the reference line.
This did not matter as these strokes were clearly parts of long
strokes crossing the line.

Proto-ogham is the simplest form and has no vowels. It comprises
various arrangements of short strokes above and below the reference
line and long strokes crossing the line, plus the special characters.

Vowels were introduced with Classic Ogham and were represented by dots
on the reference line, most commonly in the form of chips or nicks on
that edge of the stone separating the two faces. Alternatively, they
could be represented by short strokes crossing the reference line. As
previously, the angle at which any of the strokes met/crossed the
refrence line did not much matter. Short strokes could still be
distinguished from long strokes, and strokes crossing the line could
be distinguished from strokes not crossing the line.

By the time of Scholastic Ogham vowels were represented by strokes
crossing the reference line at right angles and those characters which
had always been represented by strokes crossing the reference line
were distinguished from the vowels by being carved at an angle to the
reference line. See
http://www.summerlands.com/crossroads/library/ogham.gif

To further complicate the issue, the exact nature of some of the
inscriptions is not known. In these cases there are no photographs or
accurate drawings. We do have diagrams which in at least some examples
are transliterations. Nobody knows what was really there. All we can
be certain of is that when so-and-so made a copy of it he wrote down a
transliteration of what he thought was written there. In other words,
all we have is his idea of what he thought it said.

Problems have arisen from the fact that some vowelled Oghams employ
strokes for vowels while others employ dots. Larry had read
extensively and found a number of examples where he suspected that the
transcriber had written down vowels on the basis of what may be no
more than accidental chips in the edge of the stone. In some cases
this has caused him to try to read them differently and in the
subsequent internet discussions this gave rise to accusations of his
having 'edited' the original inscription.

  ------------

Larry Athy was confronted with the problem that a considerable
proportion of the Ogham inscriptions could not be read, or could only
be read with considerable uncertainty. Not only had he studied all the
generally recognised Ogham inscriptions, and read the books written by
the recognised authorities, but he was the official collector of Ogham
information for ESOP. Consequently he probably knew of more Ogham
inscriptions than anyone else in the field.

He was concerned at the high proportion of apparently unreadable Ogham
and began to wonder if all inscriptions were written in the same
language or the same script.

What follows will make more sense if you have read the paper (PDF
available on request) but, in brief, what Larry did was to consider
the possibility that some of the unintelligible Oghams in the British
Isles (includes Ireland) may be unintelligible because they were
written using a vowelless Ogham not recognised by scholars. So he set
up a simple filter to separate inscriptions written in the possibly
vowelless Ogham.

In fact he had three groups of Ogham: the 'High Plains' [HP-1] Oghams
from North America, the 'British Isles' Oghams which could not be read
reliably [BI-1] and the 'British Isles' Oghams which could be read
reliably with agreement amongst scholars as to what they meant [BI-2].
One of Larry's criteria for his corpus BI-2 was that the inscriptions
be read in the orthodox fashion, that is not backwards or upside down
and it may be that this has excluded at least one valid Ogham
inscription. In other words, apart from the doubtful one to which I
have just referred, BI-1 contained the rejects.

Larry Athy concluded that his thinking to this point lead to the
construction of a testable hypothesis and this is what he described in
his paper. Unfortunately his paper did not set out the various steps
by means of which he got to his hypothesis and its sudden emergence
without prior warning proved shocking to some. Nevertheless it is
possible to follow the steps which led to the construction of his
hypothesis and to then follow the analytic procedures he used to test
it.

----------------------

As I have already written in a private email, Larry's thinking
developed as follows:

Begin quote
++++++++++++++++++++++++++++++++++++++++
1. That languages sharing the same alphabet have different
distribution of symbol frequencies within bodies of text is beyond
dispute. Certainly there is room to discuss the fine details but the
main point stands.

2.  If Kevin Jones is to be believed, the Book of Ballymote makes
refrence to an ancient vowelless form of the Ogham script consisting
of 15 consonants only. [1]

3. It does not matter how the ancient vowelless form of the Ogham
script is used but two bodies of inscriptions sharing the same base
language and method of usage (encoding) will show statistical
similarity with a sufficiently large sample of inscriptions. It does
not matter whether both sets do or don't encode vowels, or use the
method of Fell (or Larry), or have had their alphabetic equivalence
selected by a table of random numbers: providing they both share the
same script, language and manner of usage, they will show statistical
similarity.

4.  Larry has shown that:
   (a)  His BI-1 and BI-2 corpora are not statistically similar under
          any circumstances.
   (b)  His BI-2 and HP-1corpora are not statistically similar under
          any circumstances.
   (c)  His BI-1 and HP-1 corpora are statistically similar when both
         are read as being written with a 15 character alphabet.[2]

5.  Larry's conclusion is not dependent on his manner of reading or
transliterating the various inscriptions, as long as he uses the same
sets of rules for each. In fact, it doesn't even matter whether he
regards the inscriptions as using vowels or not. He appears to have
established a statistically significant similarity.

6,  It is not known whether the BI-1 corpus [3] employs a spoken
language or not. However, the frequency distribution of the
characters/strokes employed is not random as might be expected if the
inscriptions were merely a list of tallies etc. Further, it is a
reasonable presumption that the use of the script may have flowed on
into the inscriptions which we do know employ a language. Even if the
inscriptions are not written directly in a spoken language they may
carry a message constructed according to formal rules. If they do they
will be sufficiently structured to qualify as a language for the
purposes of statistical analysis. Accordingly the 'language' may be
characterised by the the frequency distribution of the
characters/strokes employed.

7.  Exactly the same considerations apply to the HP-1 corpus except
that we do not have the inscriptions flowing on into a known written
language. Nevertheless, the inscriptions can be statistically
characterised in the same way as with the BI-1 corpus.

8.  The remarkable thing is the statistical similarity between HP-1
and BI-1. It is easy to write this off as mere chance but the
statistics appear to say that the odds against this are high. In other
words, it appears as though the rules to construct the HP-1 corpus may
be the same as for the BI-1 corpus. If true, that is an extraordinary
thing indeed.

9.  Further, if the BI-1 corpus is held to employ a spoken language
then it is likely that the HP-1 corpus employs the same spoken
language.

==================================================
Notes:

[1]  I now understand your point about the various ways in which
vowels may be coded and Kevin Jones was clearly referring to this in
some places. The argument has become confused by equating the question
of what is meant by Ogham Consaine with the question of the forms of
vowelless Ogham. I accept that the name 'Ogham Consaine' might not
apply to all forms of 'vowelless Ogham' so I shall avoid the term
'Ogham Consaine'.

[2]  If proposition 1 of Kevin's recent post to the news group is
accepted, that there may in the past have been a use of a 15 character
Ogham, which used vowel pairs to code vowels (and which was a 'right
mess') then an uninformed reader would see only 15 different
characters.

In much the same way with our numerical system, an uninformed reader
would see the use of only 10 different characters.:

 0  1  2  3  4  5  6  7  8  9  

12, for exaple would not be seen as new character but merely the use
of two of the old characters. In fact that is how we see it ourselves.

As I understand it, Larry's readings of his BI-1 and HP-1 corpora
employ a 15 character script in this manner.

[3] A problem which has plagued all the various discussions is Larry's
transliterations of the inscriptions. Larry has not always employed
the transliterations of Macalister etc, and possibly with good reason
in many cases. Most of the data from which people now have to work is
not of the highest reliability. I believe you have copies of all of
the inscriptoins which I refer to below.

Skellig Michael is the only inscription of which there seems to be a
photograph.

There is also a technically competent sketch of Castlederg which gives
a good impression of the inscription on the stone.

Ballyhank #5 from Macalister is not too bad either in that it is a
drawing of the stone and not a transliteration.

Piedra de Gaedhil is an excellent illustration of the problems which
can arise from transliterations. While the stone is reasonably well
drawn, it is clear that the transliteration is not an entirely
faithful reproduction of what is on the stone in that long vowel
strokes are given where there are only dots (or possibly even chips)
on the edge of the stone. One would get an entirely false idea of what
was on the stone if only the transliteration was available.

Now, look at Slieve-nacaillighe from Brash. It is not clear the extent
to which this is a sketch or a transliteration. The stone is clearly
drawn with perspective but it is not clear that perspective applies to
the long strokes at each end of the line. Has Brash tried to read it
and write it down so that it conforms with what he knows?

How reliable are the tidy lines of Bweeng when one compares them with
http://www.megalithomania.com/irishgenius/zGarrane2.jpg which may show
the original stone (centre one of the three)?

It is well known that people try to interpret things in terms of what
they believe to be true. Witnesses to motor accidents tend to recall
events in terms of their own reconstruction of what they believed
happened and, along with things which they saw, they will with equal
certainty tell you things they believe would have happened in their
particular scenario.

Much the same phenomenon applies to people who are trying to decipher
an inscription. They will try to read/interpret/transliterate in the
terms of the character sets they know. If they expect vowels to be
present, they may try to find them, using even the least hint of a
line or chip as evidence that once a vowel was there. This is not
dishonesty. This is normal human behaviour.

What I understand Larry has done is go back to the inscriptions of his
BI-1 corpus and re-examine such raw data as there is, to determine how
reliable the presumed presence of vowels may actually be.
Unfortunately, I don't have all of Larry's diagrams. Nor do I have all
his sources but I do know that at various times he has come out with
comments which suggest that his decision to not read vowels is not
entirely unfounded. This is not 'editing' or tinkering with the
inscriptions. It is merely considering what is known from a different
point of view.

Unfortunately I do not have the data to enable me to confirm what he
has done in detail.
++++++++++++++++++++++++++
End quote

I'm not going to try and defend Larry Athy's theories except to say
that they were not the product of quick and dirty half-baked analysis
and he did test his hypothesis as well as he was able. His views are
different from the main stream but that does not mean that he was
wrong. The fact that they were different does not mean that they were
'garbage' and certainly GIGO as most recently implied by Doug Weller
does not necessarily apply.

Eric Stevens