03 March, 2011

Pictish language: Inscriptions

The people living in what is now Northern Scotland in the Late Roman and early medieval period were known to their Latin-speaking neighbours (and, by extension, to us) as Picts. They spoke a distinct language, but little if any of it survives. A small number of Pictish inscriptions using the Irish script called ogham are known. What can these tell us about the lost language of the Picts?



Most of the surviving Pictish inscriptions use the Ogham script, which was invented in Ireland perhaps around the fourth century and introduced to what is now mainland Scotland via Dal Riada (an Irish-speaking early medieval kingdom in roughly the area of modern Argyll, see map). Ogham is written by making straight strokes across, sloping from or perpendicular to a guideline, originally the edge of a stone.

Twenty-nine Pictish ogham inscriptions have been found, mostly on stones, with three on knife handles (Bac Mhic Connain, Hebrides; Aikerness, Orkney; Weeting, Norfolk) and one on a spindle whorl (Buckquoy, Orkney). Pictish ogham started out closely following the Irish, but later developed more ornate forms and began to be carved on a guideline on the face of the stone instead of along its edge. Most Pictish ogham inscriptions are believed to date from the seventh to tenth centuries (Laing & Laing 2001).

Where they have been deciphered the ogham inscriptions appear to record names, some P-Celtic (Brittonic), some Q-Celtic (Gaelic). Sometimes words of apparently Irish derivation appear, e.g. “meqq” which is thought to be related to Irish “maqq” meaning “son of”. One from Bressay, Shetland, contains “crroscc” thought to be the Irish word for “cross” and “dattr” which is thought to be the Norse for “daughter”.

Some have not been deciphered, for example the baffling inscription on the Lunnasting stone in Shetland, which reads:

ettocuhetts ahehhttann hccvvevv nehhtons

The last word looks recognisably like the Pictish personal name Nechtan, recorded in the Pictish king-list and by Bede. The rest of the inscription has defied all attempts at interpretation so far.


Incomprehensible inscriptions such as the one on the Lunnasting stone have led to suggestions that the Pictish language, or a sizeable component of it, was unrelated to any Celtic language or indeed any surviving language, with the further suggestion that Pictish, like Basque, could have been a survivor of a pre-Indo-European language dating back into the far distant past. However, some of the inscriptions may be incomprehensible only because we have not yet figured out how to read them. For example, the inscription on the Buckquoy spindle whorl was initially thought to be written in an unknown non-Celtic language, but can be interpreted as an intelligible inscription in Old Irish meaning “a blessing on the soul of L” (Forsyth 1995), presumably a good-luck charm of some sort for a person whose name began with L. Some of the other apparently incomprehensible Pictish inscriptions may be similar.

Another possibility may be that the undeciphered inscriptions are written in some condensed form, perhaps abbreviations or shortened forms of words that were readily understood at the time. Modern British coins carry the inscription “ELIZABETH II D G REG F D”. This is a shortened form of a Latin inscription, “ELIZABETH II DEI GRATIA REGINA FIDEI DEFENSOR”, meaning “Elizabeth the second, by the grace of god, queen and defender of the faith”. The personal name is fairly readily recognisable, but trying to reconstruct the full Latin inscription from the abbreviated version could be a challenge if the abbreviation conventions and the Latin language had been lost.

Another possibility is that the incomprehensible inscriptions were not intended to be read as a written script. There is a Pictish inscribed stone at Newton, Aberdeenshire, with an inscription carved in an otherwise unknown (and indecipherable) script, together with an ogham inscription along its edge. The ogham has been dated to the seventh or eighth century, and the main inscription is believed to be contemporary on the basis of the weathering. The main inscription resembles Irish majuscule script, and may have been carved by someone who was illiterate in that script (Laing & Laing 2001). Perhaps the carver, or the patron who commissioned the stone, thought the Irish symbols were beautiful or powerful or both, and carved them as an image rather than as written symbols with meaning. There are other examples of letters used in a way that does not convey a written message; for example, a seax (short sword) blade recovered from the River Thames was inscribed with the full 28-letter Old English runic alphabet. Presumably the runes were there for some purpose other than conveying a written message. Possibly the indecipherable Pictish ogham inscriptions were also using the ogham letters for some purpose other than writing a message.

The inscription on the Lunnasting stone contains many doubled letters. This may be consistent with the letters being used for some purpose other than writing a message. However, doubled letters are not unique to Pictish inscriptions, as English runic inscriptions sometimes contain doubled letters. For example, the Ruthwell Cross inscription contains several examples of words with doubled letters: almeittig; riicnae; gistoddu (Page 2003).

The reason for the presence of doubled letters in runic inscriptions is unknown (Page 2003, p.148). It may be pure chance, or it may reflect some tradition among the carvers of monument inscriptions, or some prosaic reason such as a simple mistake. If you find yourself mistakenly carving the letter you have just carved, you can’t rub out the mistake and correct it, and you can hardly go and make another stone monument and start all over again. It would be simpler just to finish carving the mistaken letter and then carry on with the rest of the inscription. Mistakes would be expected to be more common if the carver was working in an unfamiliar script, as may have been the case for Pictish carvers as the ogham script originated in Ireland. My thanks to Doug Tankard for providing a modern example of an incomprehensible inscription with doubled letters:

Copyright: Doug Tankard.

This is a towel manufactured in China and displaying the motto and crest of the English football team Arsenal. The motto should read ‘Victoria Concordia Crescit’, which is Latin and translates as ‘Victory comes from harmony’. The Chinese manufacturer, presumably (and quite reasonably) unfamiliar with both Latin and Arsenal Football Club, has written it as ‘Victory Contoral Crrhtty’, which contains a recognisable word (Victory) and two words that make no sense. Just imagine trying to reconstruct the language in use in 21st century England if this was the only surviving piece of evidence.

On the basis of the inscriptions, I would be cautious about concluding that Pictish was a unique non-Celtic language, although it may well have contained a non-Celtic component that distinguished Pictish from the Irish and Brittonic languages spoken elsewhere in Britain. Place names also provide a few snippets of evidence about the Pictish language; more on this in another post.

Forsyth K. The ogham-inscribed spindle-whorl from Buckquoy: evidence for the Irish language in pre-Viking Orkney? Proc Soc Antiq Scot 1995;125:677-696. Available online
Laing L, Laing J. The Picts and the Scots. Sutton, 2001, ISBN 0-7509-2873-5.
Page RI. An introduction to English runes. Boydell, 2003, ISBN 0-85115-946-X


Rick said...

I notice that the Chinese rendition of the Arsenal motto goes downhill, so to speak - a word, something that at least looks like a word, and then a sheer jumble - interesting, though, that the last has doubled letters and in fact looks a lot like the Pictish inscriptions (as Romanized).

I shudder to think of what Westerners do to Chinese characters.

Back (somewhat) to the point, I seem to recall reading that the temptation to tie any generally indecipherable European inscription to Basque is all but irresistible.

Carla said...

Rick - Indeed, one would probably try to copy the Chinese characters as an image, but without knowing the meaning would probably make all sorts of mistakes (wrong stroke length, angle, thickness....) that might render the characters illegible without realising it. Or if one had a pre-existing set of Chinese characters (sort of like a font), one would probably try to pick the closest match by eye, but again, without knowing the meaning, would probably pick the wrong match a lot of the time with incomprehensible results.
If the Chinese towel manufacturer had something like a photograph of the Arsenal motto to work from, maybe the photograph was clearer at one side than the other, which could account for the difference. Or maybe the words were easier to recognise and interpret the closer they were to modern English words, so Victoria looked similar to Victory and was manageable, but Concordat and Crescit were all at sea.

The origin and relationships of the Basque language is still something of a puzzle, and puzzles attract other puzzles :-) Perhaps in the hope that if you've got two questions you can't answer, you might be able to put them together and get a different question.

If Basque is a survivor of a pre-IE language from very early prehistoric Europe, then it's appealing to think it might represent some sort of universal early European language and therefore to look for echoes of it elsewhere.

Gabriele Campbell said...

There are echoes of an European language in a certain layer of hydronymic names (names of rivers, lakes), but those seem to represent a stage of an Indoeuropean language before it split into Italo-Latin, Celtic, Germanic, Slavic and Baltic languages. It's not related to Basque. But it is the oldest linguistic strata we can still trace.

Rick said...

Yes on all counts.

As for looking for Basque connections, for all we know pre-IE Europe was a patchquilt of languages that had mostly developed independently of each other long enough for no extensive family relationships to remain.

I'm sure I've asked this before, but forgotten the answer. What connection was there, if any, between the name of the Picts and the Latin word for 'painted?' I'll guess that the connection was made at an early date, but is probably a coincidence.

Carla said...

Gabriele - are those the names that include things like 'Don', which appears as a river name from Russia to Yorkshire? If I remember rightly, a few river names in Britain (not sure about elsewhere) are not easily interpreted as Indo-European and there's a theory that they may represent an older language group.

Rick - 'Picti' appears in Latin sources and is approximately the Latin for something like 'painted ones'. It may have been purely an invention of Latin writers and never used by the Picts themselves, or it may have been a Latinised rendering of their own tribal name. If it's a Latinised form of a tribal name perhaps based on something like the 'Pit-' place name element, its resemblance to Latin 'Picti' (painted) would be coincidental, although no doubt Latin writers happily made the most of such a colourful coincidence. See my earlier post about the Picts and their mutliplicity of names depending on the source and the language here.

Gabriele Campbell said...

Carla, yep, there are several strata of river names. Pre-Indoeuropean, the Ancient European Hydronymy (which relates the Weser river to Whisky, lol) and the ones of languages once spoken in a country (Viking names in the UK, for example). Rivers and lakes are more conservative in keeping names than settlements, so they have those older layers while all the variants of 'York' fe. only go back to historical times.

Carla said...

Gabriele - So Weser belongs to the Usk/ Esk/ Isca/ Uisge group of river names meaning something like 'water', yes? Hence Gaelic 'usige beatha', 'water of life', Anglicised to 'whisky'. Hard to say whether some of the other place names might have been in use long before they were first recorded and so whether they might have roots extending back into prehistory. Eboracum is usually interpreted as a Latinised form of a British name with a root related to modern Scottish Gaelic 'iubhair' meaning 'yew tree' (there's a Sgurr an Iubhair, 'peak of the yew tree' in the Mamores mountain range in Scotland. If so it would predate the Romans and there's no way of knowing how far back it goes. Some settlement names might be even older; the origin of the name London still seems to be uncertain (as far as I know) and might be pre-IE, or at any rate non-IE, like the river names.

Rick said...

I knew that river names tended to persist, but not that they showed some common features across Europe.

Carla said...

Rick - a few river names share common elements, e.g. River Don (Russia), River Danube, River Don (Yorkshire) are all supposed to derive from a common IE root word something like 'danu', which might mean 'river' or refer to a river goddess, or both. How many such roots are common to river names across Europe/Eurasia I don't know; it needs more linguistic expertise than I have to differentiate between an actual connection and a coincidence of syllables that happen to look or sound similar.