The language of the Picts


by Paul Kavanagh

Myth, misunderstanding and legend mix with fact in popular understanding of the Picts.  Many years ago I met a guy who insisted that the Picts were in fact pygmies from Africa.  He was deadly serious, though how a population of Central African people made it to Iron Age Scotland was never explained, CalMac ferries still don’t go to the Congo.  Others have maintained that the Picts spoke a language related to Finnish or Basque, or even that they spoke a Germanic language ancestral to modern Scots.   None of these theories is supported by the linguistic evidence.

These theories are usually supported by the supposed phonetic resemblances of Pictish names to words in modern languages.  Using the same methods as the proponents of Germanic or Finnish speaking Picts, we could just as easily “prove” that the Picts spoke the Nahua language of the Aztecs, Mongolian, or any language you care to mention.  What historical linguists seek to identify are repeating and regular patterns of correspondence.  Often these correspondences do not resemble one another phonetically at all. (1)  Using the tried and tested methods of historical linguistics, Celtic is the only language family that can be identified amongst the Picts.

Pictish, or pre-Pictish, names which are assuredly Celtic include the first Scottish person whose name is recorded in history, Calgacus the chieftain of the Caledonians whose speech was recorded by the Roman historian Tacitus.  The name Calgacus is from a Celtic root meaning ‘swordsman’.  It’s etymologically related to the Gaelic word calgach, with the same meaning.  The older Irish name of the city of Derry was Doire Calgaich ‘the grove of the swordsman’.

Other Celtic names include the names of some of the tribes of Northern Scotland.  The Caereni of Sutherland took their name from the Celtic word *kaeros ‘sheep’, which survives in Gaelic as caora.  Caereni meant something like ‘the shepherds’.  The tribal name Caledonii seemingly contains the Celtic word *kaletos ‘hard’ and meant something like ‘the tough ones’.

No Pictish literature survives.  We have no Pictish manuscripts, and what little that is known about the Picts and their language must be pieced together from second hand sources in Latin, Old Irish, Old English and Old Welsh texts and chronicles, and what can be gleaned from the careful examination of place names and personal names.  Because so little hard information survives about the Picts, speculation, ideology and sheer wishful thinking rush in to fill the gap.  The Picts have been especially prone to mythologising.  As so often in history, the amount of speculation is in inverse proportion to the amount of information available.  

Many academics and scholars once held that the Picts were a pre-Celtic people with some distinctly un-Celtic habits like tracing descent through the female line.  It was also believed that at least some Picts spoke an ancient pre-Celtic tongue descended from the language of Bronze Age times.  This theory has fallen out of favour in more recent decades and these days most serious academics hold that the Picts were essentially Celtic in language and culture, although elements of aboriginal pre-Celtic culture are likely to have survived amongst them to some extent.

The majority academic view nowadays is that Picts spoke a Celtic language derived from the same Brittonic tongue which was ancestral to Welsh, Cornish and Breton.  The close relationship between Pictish and Brittonic can be seen in Pictish place names like Aberdeen and Abernethy which contain aber, the same word for ‘rivermouth’ or ‘confluence’ familiar from Welsh place names like Aberystwyth or Aberdare.  Other Pictish place names which look decidedly Welsh are Perth, Welsh perth ‘hedge’, and Gordoun, from older gor-din ‘superior, upper fort’.  In this name, as so often in Pictish names, part of the name has been replaced by a Gaelic form, Gaelic dùn replaced Pictish din.

Pictish names are usually identified on geographical grounds, as the linguistic content of these names is generally identical to Brittonic Cumbric names from south of the Forth.  Only the merest handful of place name elements can be identified as specifically Pictish.  Perhaps the best known are names beginning Pit-, which apparently derives from a Pictish word meaning ‘piece of land, farm’.  Pit names are found over a large area of Scotland north of the Forth, and their distribution correlates well with the extent of former Pictland.  However we can’t be certain that all pit- names were created by the Picts as the word was borrowed into Scottish Gaelic.  Some pit- names will have been created later by Gaelic speakers in former Pictish territory.

Despite the phonetic similarity between “pit” and “Pict” the two words are not etymologically related.  Although seemingly absent from Brittonic, the word pit- was found in the ancient Gaulish language of classical times.  The Gaulish word was pettia, believed to mean ‘a tract of land’.  Pettia was borrowed into the Latin of Gaul, the ancestor of modern French.  After various phonetic changes, it survives as the French word pièce, later borrowed into English as ‘piece’.  When you ask for a piece and jam you’re using a word that links us with our Pictish ancestors.

Gaulish was very closely related to Brittonic, in fact the two were probably mutually intelligible dialects of a single linguistic system.  Roman sources tell us that there was much traffic between Gaul and Britain, and that the Britons and Gauls understood one another without diffculty.  It seems that Pictish derived from a northern Celtic variety which, like Gaulish and Brittonic, was a part of a large Celtic dialect complex.  Originally Pictish was simply the northernmost extension of this immense Celtic dialect group, which once stretched all the way from Scotland and Ireland to Turkey. (2)  

After the Roman conquest of southern Britain, the unconquered Britons in the north went their own way culturally and politically.  By the 4th century the people of this region had come to be known as Picti by their Romanised neighbours.  

We don’t know what the Picts called themselves.  Many acres of scholarly ink have been spilled discussing whether the name Picti was simply a Latin word meaning “the painted ones”, a reference to the Celtic habit of tattooing and body painting, or whether it was an indigenous Celtic name.  One of the tribes of Celtic Gaul was known by the very similar name Pictavi.  This tribe lived near the mouth of the Loire, their name survives in the name of French city Poitiers and the region of Poitou.

The Irish and Welsh called the Picts by names deriving from the ancient Celtic word *Qritani, which evolved into Cruithne in Old Irish, and Prydyn in Old Welsh.  The same Celtic word also lies behind the name Britain, deriving from a Latin borrowing of the older Brittonic version of the name, which would have been something like *Pritani.  Latin speakers often confused the Celtic sequence pr- with Latin br-, and the early Romans misheard Pritani as Brittani.  The Romanised Celts of southern Britain later adopted this Latinised name for themselves, and retained the Celtic pronunciation to refer to the tribes north of the Forth who resisted Roman rule.  

At the beginning of the Roman occupation, the languages later known as Pictish and Brittonic were simply regional varieties of the same P-Celtic language.  By the time of the Old English cleric and historian the Venerable Bede, who lived in the 9th century, Pictish had come to be regarded as a different language from the Old Welsh of the Britons who had come under Roman rule.  Bede tells us that in his day the languages of Britain were English, Welsh, Latin, Gaelic and Pictish.  Unfortunately it is unknown precisely how Pictish differed, as it must have differed, from Old Welsh.

However we can reasonably guess that unlike the Britons, who adopted thousands of Latin words – many of which still survive in Welsh, Cornish and Breton – the Picts escaped this intense Latin influence. (3)  They would also have been isolated from the phonetic and grammatical changes which Brittonic experienced as it evolved into Old Welsh and Old Cornish, and Pictish underwent phonetic and grammatical changes of its own.  Unfortunately we don’t have details about what these changes were.  

It seems that in Roman Britain, Latin was in the process of replacing Brittonic – just as Latin replaced the indigenous languages of most other parts of the western half of the Empire – but this process was interrupted by the fall of the Roman Empire.   There are three modern language groups which descend from languages spoken by peoples conquered by Rome.  As well as Welsh and its close relatives Cornish and Breton, Basque and Albanian also descend from languages spoken by subject peoples of Rome.  All three were spoken in remote areas of the Empire where the process of Latinisation had not run its course before the end of the Empire, so these languages survived, unlike Gaulish, Iberian, Etruscan and many others which were replaced by the spoken Latin from which the modern French, Spanish and other Romance languages descend.  The Brittonic languages, Basque and Albanian are all characterised by massive borrowing from the Latin of the Empire, even words of basic vocabulary in these languages were replaced by loans from Latin.  

Modern languages descended from languages spoken beyond the frontiers of the Empire, like Gaelic or English, do not display this early Latin influence.  These languages typically have only a handful of early Latin words most commonly referring to items of Roman culture and trade goods.  The Latin words in Archaic Irish or North Sea Germanic, the Roman era ancestors of Gaelic and English, were additions to the vocabulary, they did not replace native words.  Both Gaelic and English later borrowed much vocabulary from Latin, but historical linguists are able to determine that these loanwords are later in date, and were borrowed after the end of the Empire.

We know that Gaelic eventually replaced Pictish throughout almost all of its territory.  Prior to its replacement, Pictish must have come under strong linguistic influence from older forms of Gaelic, just as the Brittonic spoken further south came under strong Latin influence.  When one language replaces another, as Gaelic replaced Pictish, the earlier language invariably comes under strong influence from the language that ultimately displaces it.  It would not be unfair to say that Old Welsh was a Romanised form of Brittonic, while Pictish was a Gaelicised form of Brittonic.

Pictish has left few identifiable traces in Gaelic.  Again this is normal in language replacement.  When languages are in contact, borrowing is not symmetrical, the language of higher prestige always influences the language of lower prestige far more than the lower prestige language influences the high prestige language.  The language doing the replacing typically shows relatively little influence from the older tongue.  We see this in modern Scotland, where Standard English has largely replaced Gaelic and Scots.  Modern Gaelic and Scots display massive influence from English and contain hundreds of English words, but Standard Scottish English has not been influenced by Gaelic or Scots to anything like the same extent.

Just four modern Gaelic words are thought to be loanwords from Pictish, although they could just as easily be loanwords from the closely related Old Welsh once spoken south of the Forth.  The words are dail ‘haugh, pasture land beside a river’ from Old Welsh or Pictish dôl, monadh ‘moor, mountain’ from monid (modern Welsh mynydd), pòr ‘grain’ from pawr, and preas ‘thicket’ from pres or prys.  A few other words may be of Pictish or Old Welsh origin, such as bad ‘cluster, clump’, but the status of these words as loans is uncertain.  A couple of other Gaelic words, such as peit from Pictish pit, and obar from aber, are found only in place names.

Pictish, or perhaps the closely related Old Welsh, has left traces in Scottish Gaelic in a more subtle way.  The verbal system of Scottish Gaelic is quite distinct from that of many Irish dialects.  In particular, the modern future tense of Scottish Gaelic descends from the original present tense of Old Irish, preserved as a present tense in most modern Irish dialects.  The exact same change occurred in Old Welsh, the modern future tense of Welsh descends from the present tense of Brittonic verbs.

Apart from this linguistic archaeology, we have only the names of Pictish individuals preserved in manuscripts written in other languages.  The Pictish king lists have generated much discussion but unfortunately these documents have been repeatedly copied by scribes who did not understand the source material, and are so corrupt that they are useless as evidence for the linguistic affiliations of Pictish.

A recent theory claims that the symbols found on Pictish symbol stones is a form of ‘hieroglyphic’ writing.  This may very well be true, but we are unable to give any phonetic content to the symbols, so they cannot help us determine much about the Pictish language.

However we do have a few brief and precious examples of Pictish writing which can be given phonetic form.  There are a handful of Pictish Christian inscriptions in the Latin alphabet, but perhaps the most mysterious monuments the Picts have left us are a series of inscriptions written in an ancient Irish alphabet called ogam. These enigmatic inscriptions deserve to be better known amongst modern Scots, they represent some of the oldest examples of people in Scotland leaving their own messages for posterity, instead of being reported in third party accounts.  Approximately 30 in number, the Pictish ogams are also the only first hand linguistic remains of Pictish and so are central to any discussion on their language.  The meaning and significance of the inscriptions has been hotly debated, and they will doubtless continue to be a source of disagreement.  

The ogam (sometimes spelled ogham) alphabet appears strikingly different from our Latin alphabet.  Ogam letters, called feda, consist of lines and notches inscribed into stone or other hard materials, cutting across a central stem line.  In older inscriptions, the letters were cut into the edges of stones, using the edge of the stone as the stem line.  In later ogam inscriptions, a stem line was engraved on the face of the stone.  Most of the Pictish ogams are of this later type, often called ‘scholastic ogam’.  
Unfortunately ogam inscriptions are very prone to damage from erosion and wear.  Slight damage to one part of an ogam feda can turn it into a different letter.  Reading the inscriptions can be a challenging task at the best of times and many ogam inscriptions remain unintelligible, even those in Ireland where it is known that they are written in an older form of Gaelic.

Although there have been many theories about the origins of ogams, the majority opinion nowadays holds that they were probably devised somewhere in southern Ireland in the later centuries of the Roman occupation of southern Britain.  Although it does not resemble the Latin alphabet, many scholars have argued that whoever invented ogam must have been familiar with the writings of Latin and Greek grammarians.  

It is possible, although less certain, that ogams were first used by the earliest Christian communities in Ireland, which predated the time of St Patrick who is thought to have lived in the 5th century.  According to this theory, the use of ogams spread with the early Irish Christian church, which became highly influential in Pictland.

The oldest ogam inscriptions are found in the Irish province of Munster, which seems to have been the original centre out of which the ogam tradition expanded.  Over one third of the approximately 400 surviving inscriptions come from the Irish county of Kerry, and most of these come from the territory of the ancient Gaelic kingdom of Corcu Duibhne.  The oldest ogams are dated to around the 4th century.

All ogam inscriptions in Ireland are in a very early form of Gaelic, which philologists refer to as Archaic Irish.  Typically the inscriptions are short and consist solely of personal names in fixed formulas which often contain the words MAQI, an early form of Gaelic mac ‘son’, AVI an early form of ogha ‘grandchild’, or MUCOI a word meaning ‘tribe’ or ‘clan’ which has not survived into modern Gaelic or Irish.  They are believed to be funerary inscriptions, although most are not found in association with burials.

With the spread of Goidelic culture and language in the sub-Roman period, ogam inscriptions start to appear in Scotland, Wales, and in the South West of England, all areas where early Gaelic speakers settled as the Romans lost control of their British province.  The vast majority of these inscriptions are also in Archaic Irish.  A typical example of an Archaic Irish ogam, from the Isle of Gigha in the Kingdom of Dalriata:

Viqula son of Cugini

In the north and north west of Scotland, particularly in the Northern Isles, some two dozen Pictish ogam inscriptions survive.  They are the only known examples of the ogam alphabet being used to write a language other than Archaic Irish.  Pictured at right is the Brandsbutt stone from Aberdeenshire, with its ogam inscription visible. Although they can be read, albeit with difficulty, these inscriptions defy interpretation.  The Brandsbutt inscription reads IRATADDOARENS, which may contain the name of the Celtic saint Eddaron or Etharnus.  The Pictish ogam inscriptions are believed to date from the 6th to the 9th century, when the Picts were under heavy cultural and political influence from their Gaelic speaking neighbours, who by this time were well established on the west coast of Scotland.

It is unknown how the Picts adapted ogams to suit their own language.  Ogams did not have letters for sounds not found in Archaic Irish, so for example there was no ogam letter P (4), a sound which the oldest Gaelic lacked but which was present in Pictish.  The ogam letter sequence MAQQ appears in a number of Pictish ogam inscriptions, but it’s not known whether this was simply the Archaic Irish word MAQQ (modern Gaelic mac ‘son’, Welsh map) or whether the Picts were using the ogam letters QQ to represent Pictish P.  The Celtic personal name Nechtan appears in some Pictish inscriptions as NEHHTONN although it’s known from place names that Pictish developed the ancient Brittonic sound sequence cht, preserved in Old Irish, into th as in Welsh – as occurs in the place name Abernethy.  It’s unknown whether HHT in Pictish ogams represented cht or whether it was a spelling convention for writing the Pictish sound th.  Doubled letters are very common in the Pictish ogams, but it’s unclear whether this was a spelling rule, and if so what it represented, or whether the letters were doubled simply for stylistic reasons.

The immense difficulties in reading the Pictish ogams can be illustrated by an inscription from Buckquoy in Orkney, shown left.  Inscribed on a whorl-stone, used as a weight on a loom to add tension to the fabric as it was being woven, the inscription is circular in shape.  Read off as ETMIQAVSALLC the inscription was believed for many years to be unintelligible and decidedly non-Celtic.  It was even cited as evidence for a non-Celtic Pictish language.

However in the 1990s it was found that older scholars had been reading the ogams backwards.  When read in the opposite direction, the Buckquoy inscription reads BENDACCTANIML, an Archaic Irish Christian formula which means “a blessing on the soul of L.”  In modern Scottish Gaelic this would be written beannachd anam L.

One of the implications of this new reading is that it shows there was a knowledge of Gaelic amongst the Picts of Orkney before the arrival of the Norse in the islands in the 800s.  This certainly doesn’t mean they were all native Gaelic speakers, but it does demonstrate that the Picts of Orkney had a knowledge of Gaelic and that Gaelic was culturally prestigious amongst them.

Other Pictish ogams contain recognisable words or names, but the rest is unclear.  A Pictish inscription from Aboyne in Aberdeenshire reads:


The Gaelic element MAQQ is clear here, the last part TALLUORRH seems to be the personal name Talorc, and the first part of the inscription may be the personal name Nechtan, but the remainder of the inscription defies interpretation.

The Old English scholar Bede gives us the valuable information that the Pictish name for Kinneil on the Firth of Forth was Peanfahel.  The modern name Kinneil comes down to us from the Gaelic Ceann Fhàil.  A Welsh chronicle informs us that the Old Welsh name was Pengual.  All three mean “the end of the wall”.  Kinneil is of course situated at the eastern end of the Roman Antonine Wall.  The fahel, gual, fhàil element in this name derives from the Latin loanword vallum ‘wall’ (also the source of the English word).  In Latin, the letter V was pronounced W and the word was borrowed into Celtic as *wallo-.  

Languages change over time and the precise sound changes are specific to individual languages.  In Gaelic the older w sound of Celtic ended up as f, whereas in Welsh it became g or gw.  It’s known from place names north of the Forth that the Picts also developed an original w sound into g, just as in Old Welsh.  The place name Gordoun is from older *wor-din, the Gaelic equivalent is preserved in the name Fordoun.  What is most interesting about the Pictish name Peanfahel is that it apparently shows the Gaelic treatment of older w as f.  It seems that word ‘fahel’ in this name – the h was probably silent – was borrowed by the Picts from Gaelic.  

Together, this slight body of evidence suggests that the Picts spoke a Celtic language closely related to Old Welsh, but one which was under strong Gaelic influence.  Possibly this Goidelic influence was present in Pictish from the very formation of the Picts as a distinct people.  They remained under strong influence from Gaelic throughout their history, until eventually Gaelic entirely displaced Pictish.  In the Northern Isles and Caithness, Pictish most likely gave way to the Old Norse of the Vikings.  

Pictish died out unlamented and unrecorded.  It had already started to give way to Gaelic before the demise of the Pictish kingdom and its union with Dalriada to form the Kingdom of Scotland.  These late varieties of Pictish must have been strongly marked by influence from Gaelic, and possibly clung on in remote regions for a few generations longer.  By 1000 at the very latest, Pictish breathed its last.

1.  This is not the place to get into the methodology of historical linguistics, so a simple example must suffice. The sounds f and d do not resemble one another, yet we can identify a regular correspondence between f and d in some accents of English and German.  F in Cockney fin, free, fink corresponds regularly to German d in dünn, drei, denken.  Standard English and Scots have th in these words, thin, three, think.  In Norwegian, we find a t in the corresponding position in these words, tyn, tre, tenke. Th, sometimes written with the special letter þ, was the sound in the proto-Germanic, the common ancestor of Standard English, Cockney English, Scots, Norwegian and German.  

2. The language of the Galatians of the New Testament was at the other end of this dialect complex.  According to Greek sources the Galatians descended from Gaulish tribes who settled in central Anatolia in modern Turkey in the 3rd century BC.  Although they were under immense cultural influence from the Greeks, the Galatians tenaciously clung on to their Celtic tongue.  It is last heard of in the 5th century, when St Euthymius tells us of a Galatian monk who was possessed by a “demon” and as a result he found himself unable to speak Greek, only his native Galatian.  This is probably a description of a stroke suffered by the unfortunate monk.  The Galatian language had most likely died out by the 7th century.

3.  The Latin influence upon Brittonic was massive.  Approximately 800 words in modern Welsh are believed to derive from loanwords taken from Latin during Roman times.  Even basic vocabulary was replaced by Latin loanwords in Brittonic, although they’ve been much altered by later phonetic change and no longer closely resemble Latin.  Welsh words descended from Latin words borrowed during Roman times include:

anifail ‘animal’, Latin animalus
asgell ‘wing’, Latin axilla
braich ‘arm’, Latin bracchium
coch ‘red’, Latin cocchium
gwyrdd ‘green’, Latin viridum
pluen ‘feather’, Latin pluma
pysgod ‘fish’, Latin pescatum

Sc. Gaelic and Irish, descended from a language spoken outside the Roman Empire, have native Celtic words for all these concepts.  For example, ‘fish’ in Gaelic is iasg, directly descended from the original Celtic word *êskos.  Pictish probably preserved the native Celtic terms here too, but we cannot be certain how the Picts pronounced them.

4.  In later centuries, Irish writers devised additional ogam letters, including a letter for P.  These additional ogam letters, called forfeda ‘extra fedas’, are not found in the oldest inscriptions.

It is in truth not for glory, nor riches, nor honours that we are fighting, but for freedom – for that alone, which no honest man gives up but with life itself.