A Consideration of Pictish Names

by Tangwystyl verch Morgant Glasvryn
copyright 1996 Heather Rose Jones, all rights reserved

II. The Material

A. Classical Roman Sources

1. The Ptolemeic Map

The earliest tribal and place names recorded for northern Britain come from the geographical writings of the Greek Claudius Ptolemaeus in the mid-2nd century AD. My principal source of these names (Jones & Mattingly) gives them in Latinized forms, but Jackson (in Wainwright) supplies some of the Greek forms. Ptolemy's most likely source of information would have been the Flavian campaigns from approximately the last quarter of the 1st century. In Jones & Mattingly's "An Atlas of Roman Britain" they have been organized on a map with place names associated with the tribal name from the same area. Here is a summary of the information. The main headings are tribal names; the indented items are place names that appear to fall in the indicated tribal area. The interpretations are largely from Jackson. The key to the forms is:

What we find here is a mixture of Celtic and not-provably-Celtic names existing side by side in the same tribal territory, although there is a slight trend away from clearly Celtic names toward the east, considered to be the "Pictish Heartland".

2. Other Classical References

In addition to the Ptolemeic names, there are several recorded personal names of rulers from the north of Britain. Tacitus (in A.D. 97) mentions the Caledonian leader Calgacus. While Jackson interprets "Calgacus" as a Celtic name, he notes that "Caledonia" (which the early Romans used for all of Scotland) is not provably Celtic. Dio Cassius (A.D. 210) mentions a Caledonian "Argentocoxos" which is clearly a Celtic name meaning "silver- limb". There is an inscription from ca. A.D. 230 found at Colchester reading in part "LOSSIO . VEDA . DE . SVO . POSVIT . NEPOS . VEPOGENI . CALEDO", i.e. "[dedicated by] Lossio Ueda, at his own cost, nephew/grandson of Uepogenus, Caledonian". The names appear to be Celtic -- one of the names in the later king- lists (Uipoig namet) may contain a version or distortion of Uepogenus. The formula, highly unusual for a Latin inscription in giving a relationship other than "son of", is discussed in more detail later. Jackson speculates that "Lossio" may be related to the root in the place name "Loxa", interpreted as "crooked", and that "Ueda" may reflect the Celtic stem "*weid-" meaning "know". (These are clearly speculations, however.)

Of particular interest is the mention by Diodorus Siculus in the 1st century B.C. of the "Pretanic" islands (i.e. Britain), deriving from a tribal name of *Pritani or *Preteni. (Wainwright p.135) Although the name itself appears to be Celtic, this tribal name was later closely associated with the Picts. In medieval times, the Welsh used "Prydain" to name Britain and "Prydyn" for "Pict" or "Pictland". The medieval Irish borrowed the word and adapted it to their own sound-system, using "Cruithin" (plural "Cruithni") to refer not only to the Picts but to British immigrants in Ireland. See the discussion in section II.B.1. for further aspects.

One of the things strongly suggested by the early data is that, whether or not a non-Celtic Pictish language still flourished by the 1st century A.D., place names and personal names of people associated with the Pictish strongholds show a mixture of Celtic and non-Celtic elements at this time. From a practical point of view for the purposes of naming, this means that a Celtic name would be appropriate for a Pict to bear. It probably means that a Celtic language context would be appropriate for a Pictish name at this time. And we have evidence, at least from the 3rd century, of a "Caledonian" name using native elements (and probably a native relationship formula) being rendered in Latin in a memorial inscription.

B. The Pictish Chronicle

The majority of names associated with the Picts are found in a collection of chronicles and histories originating in Scotland. There are scattered names other sources, but the primary one is a document called the "Pictish Chronicle" giving the pseudo-history version of Pictish origins and several lists of kings. The kings involved (if you take the chronology of the manuscripts at face value) date from several centuries B.C. through the mid-9th century. However, the earliest manuscript copy of the list dates only from the 14th century (the "Colbertine MS"). There are many other later versions, but they show so many copying errors and corruptions that I will omit consideration of them except for some narrow comparative purposes.

The primary comparison I will make is with an Irish translation of the "Historia Britonum" that was made in the 11th century (attributed to a man named Gillacaemhin), with the earliest copy appearing in the "Leabhar na h-uidhre" compiled by Maelmure who died in 1106. Only a fragment of this is preserved, however, and the earliest complete copy is found in the late 14th century "Book of Ballimote", where the Pictish material is found. Curiously, in many cases, this Irish manuscript appears to preserve older forms of the names than the Scottish one. A great deal of confusion is introduced in the Irish manuscripts because they evidently were copied from an original containing two columns -- the list of "Brude"s and the remainder of the text -- and this was not understood by the copyist who treated the whole as a single column. Fortunately, other sources make this fairly easy to untangle, however it shows how easily data can be mangled beyond recognition when copied by someone who does not understand the original material. (Skene p.xxxiff) Unless otherwise specified, names will be presented in two columns with the Colbertine form first and the Irish form second.

The Chronicle contains a preface that contains passages extracted from the 7th century "Origines" of Isidore of Seville mentioned earlier, a list of Pictish kings that can be divided roughly into "the eponymous kings", "the Brudes", and "the historical kings", and a third section concerning the Scottish kings from Kenneth MacAlpin onward. The original probably had an additional section before this last describing the destruction of the Picts and the "treacherous slaughter of their nobles at a meeting with the Scots", for a summary of these events is found in that place in Ranulph Higden's 14th century "Polichronicon" in a section clearly derived from the Pictish Chronicle. (Skene p.xixf) The date of this slaughter is placed around 850 (Skene p.cxci) which pretty much marks the close of the Picts as a documented ethnic or political entity.

The second and third sections of the chronicle give evidence of having been translated from Gaelic into Latin by someone not entirely comfortable with Gaelic. For example, the phrase "Dadrest" has in one case been interpreted as a personal name, while the context (and a knowledge of Gaelic) clearly shows it as the phrase "two Drests". There are other similar occurances. As discussed below, there appear to be several other layers of misunderstanding that have been incorporated into the king-lists themselves, with the result of greatly confusing an understanding of the names involved. (Skene p.xxif)

1. Cruithne and His Sons - the "Eponymous Kings"

The first section of the king-lists focuses on "Cruithne" ("Cruidne" as it appears in the manuscript) and his seven sons, for whom the Pictish race and the seven principal divisions of Pictish territory were said to be named. This follows the standard format for pseudo-historical origin myths and should not be interpreted as evidence for actual personal names.

The name "Cruithne" demonstrates one of the major strains of difficulty in interpreting the written evidence. It is a Q-Celtic (Gaelic) version of the tribal name "Pritani" mentioned above. Up until the period of massive change in the P-Celtic and Q-Celtic languages that occurred around the 6th century, there is evidence that people were widely aware of the major sound-correspondences between the two languages, and that words and names were often "translated" between them as if producing a cognate derived from an earlier common root. Thus the Latin name "Patricius" (Patick) appears in some cases in Ireland as "Cothric", substituting "c" for "p" as if "Patrick" had been a Brythonic name for which there might reasonably be a Goidelic cognate. Similarly, if we assume "Pritani" (or something similar) as the original tribal name involved here, we can understand "Cruithne" as a Gaelic "translation" of the name. A comparison of Scottish and Irish versions of the Pictish king-lists shows many other examples of this sort of "translation" and points out the difficulty in taking manuscripts at face value.

The names beginning in "f" also show evidence of having Q-Celtic origins. (Original Celtic initial "U-" became "F-" in Q-Celtic and "Gw- in P-Celtic, whereas P-Celtic has no native words beginning with the sound [f].) While it is possible to entertain an explanation that these names represent non-Celtic Pictish forms, a far more likely answer is that they are Q-Celtic placenames, dating from after the Gaelic influx, that have had a false antiquity lent to them through this origin myth. Alternately, they may be Gaelicizations of names that _do_ belong to some valid early tradition. "Fib" may correspond to the "Uip" found in the list of Brudes and possibly also to the "Uepo-" in "Uepogenus".

All of these factors lead one to the conclusion that the development of the "origin-legend" of the Picts involving Cruithne and his sons did not occur until the region was primarily Goedelic rather than Brythonic-speaking. And this, in turn, brings one to the conclusion that these figures are fictional and that their names do not necessarily reflect actual naming practices.

Cruithne is given only a father in the Colbertine MS (Cruidne filius Cinge), but in the Irish version of the "Historia Britonum" he is given the following pedigree, tracing all the way back to the Biblical Noah: Cruithne mac Cinge mic Luchtain mic Parrthalan mic Agnoinn mic Buain mic Mais mic Fathecht mic Jafeth mic Noe.

The majority of the names of the seven "sons" can be identified with the names of known regions in Scotland, as indicated below.

FibFib"Fife"
Fida(i)chFidach?
FloclaidFodlaappearsin"Atfodla",theoldformofAthole
FortrennFortrendwesternpartofPerth
GotCaitappearsin"Caithness"
CeCe?
Circin(n)Cirighappearsin"Maghcircin">"Mearns"
After the seven "sons" there are six other names in this section (together with the length of their alleged reigns -- 100 years in one case!).

It is unclear whether this information is at all reliable. Several of the forms appear to be Gaelicized ("Fuidid", endings in "-ach") while others are P-Celtic forms ("Wurgest", "gurcich"). The internal chronology suggests that any Q-Celtic forms must be incorrect. Interestingly, the Irish manuscript has preserved at least one P-Celtic form ("Guidid") where the other has a Gaelic form. For a discussion of a phenomenon in the last two names, see the next section. In general, I would not consider these names reliable, at least for the period they claim to represent.

2. The Thirty "Brude"s

The second part of the king-lists contains a list referred to as the "thirty Brudes". In fact, the available lists contain only twenty-eight, unless you count "Gest" and "Urgest" from the end of the preceding section (which would certainly fit the pattern). There is also a small number of other names following this. I will treat the two sections separately. The best way to begin consideration of the Brudes is simply to present the data. In this case, the Scottish and Irish lists are highly similar (pointing, perhaps, to a stronger reliance on manuscript forms, or simply to an acknowledgement that the names were alien and could not be "translated").

It is not linguistics, but logic, that provides the first analysis. We have a sequence of symbolic elements of the form: XA, XA', XB, XB', XC, XC', ... Xi, Xi'. Something funny is going on here, and we should be very suspicious of taking the data at face value. Skene makes no critical analysis of the structure of this list, other than to note that the list is said to contain "the names of the men, and the portions of the men" and to suggest of the second elements of the entries, "That these monosyllables enter into the composition of the Pictish proper names is plain enough; but they probably also entered into the names of smaller districts, which cannot now be identified." In other words, he seems to be interpreting the second elements as locatives -- even while noting that some of the elements in question are found elsewhere as personal names -- and taking "Brude" as the given name of thirty Pictish kings in a row. (Skene p.cv) I find this highly doubtful.

Jackson, disapointingly, makes no comment on the phenomenon, confining himself to discussing the linguistic forms and not their functions. There is a temptation to conclude that if thirty (or twenty-eight) Pictish kings in a row are identified with the word "Brude", that perhaps "brude" is simply the word for king in Pictish. There is a hint that Chadwick agrees with this theory, for discussions of the "thirty Brudes" are referenced in the index under "Brude, as a title" -- but the text itself makes no explicit commentary on the subject that I can find. The word also appears as a name later in the lists (as "Bridei", et al.), but this would not argue conclusively against the previous theory. There are numerous examples in early Welsh genealogies of previous "titles" (usually Latin ones) being used as given names. This may be a misinterpretation by later copyists, or it may have been the actual adoption of the element as a personal name. The use of "Brendan" as a given name in Irish is a more certain parallel. It was borrowed from a Brythonic word for king ("brigantinos") and its removal from the context in which it was used as an ordinary word no doubt made it easier for it to be used as a personal name. Similarly, if "brude" were a common noun in Pictish, but knowledge of that language were being lost, it might contribute both to the reinterpretation of the early king- lists as using it as a personal name, but perhaps also to the adoption of it _as_ a personal name by later generations. Whatever the true explanation, there is reason to be suspicious of considering "Brude" as a personal name for the alleged era of this part of the list.

The second phenomenon we see can be interpreted with with a bit more certainty. Sequences of personal names doubled with the same name prefixed by "Ur-" or "Gwor-" (a later version of the same element) can be seen in at least one early Welsh genealogy. Bartrum (p.125 note) points out a similar pattern in the genealogy of the kings of Gwynedd given in Harleian MS 3859: "... Cein map Guorcein map Doli map Guordoli map Dumn map Guordumn ...." Chadwick (pp.3-5) discusses the phenomenon in depth, also pointing out Irish parallels, but fails to draw any clear conclusions other than that something suspicious is going on. I have a memory of reading one promising hypothesis on the subject (which follows), but unfortunately can no longer remember or relocate the source. So the following will have to be put forth on my own with appropriate credit to the unknown source.

Although the element is unquestionably found in personal names, from Gaulish "Vercingetorix", to Welsh "Gwrgant", to Irish "Fergus", it is also a preposition meaning "over, above, on". The name "ghosts" that appear in the Pictish and Welsh genealogies seem likely to be remnants of an oral formula where the preposition is used to mean "after" (or possibly "before", in which case the interpreted order may be erroneous). Thus, removing the "Brude"s and interpreting the "ur" in this light, we get "Pant, after Pant, Leo, after Leo, Gant, after Gant, Gnith, after Gnith, Fecir, etc." If the name begins with a liquid sound, as in "Leo" and "Ru", the "r" of the preposition is assimilated to the following sound: "Uleo", "Eru". This is a fairly common linguistic phenomenon and presents no problems in interpretation. (Compare, in English, with the fate of the prefix "in-" meaning "not" in "illegal" and "irregular".) If the preposition should instead be read as "before", we get instead, "Pant, before Pant, Leo, before Leo, Gant, etc." If this were a common formula at some point for reciting genealogies or regnal lists, it might have continued to exist either in oral or written form past when the formula was understood, and then was reinterpreted as including the "ghost" duplicate names.

The strongest argument against this theory is that nowhere else in Celtic languages do we find clear evidence of a cognate of "Ver-" being used to describe relative position in time, rather than space.

3. The Post-Brude, Pre-Historical Kings

Following the Brudes in the same section of the manuscript is a list of kings that lack the "Brude (Ur)-X" format, but that are separated by text from the list containing names that can be corroborated by other historic documents. In content, they would seem to belong with the "historic" kings, but for format's sake I have considered them separately. In this list, we begin to see a number of the names that seem popular in the "historic" lists: Talorc, Drust(an), Nectan -- although, interestingly, no versions of Brude. We also see several names that are clearly not Celtic in origin and rather unusual in appearance: Bliesblituth, Usconbuts, and Canutulachama. (Wainwright p.145) In this, as in the "historic" list, the names appear to be a mixture of clearly non-Goedelic Celtic names, such as Drostan, Uuen, Tarain, Lutrin, Onuist, and Uurguist; of Celtic names that are not distinguishable in their appearance between Goedelic and Brythonic, such as Talorc(an); and of names that cannot confidently be identified as Celtic (to the above, Jackson adds Bredei, Derelei, and Bargoit). In this "post-Brude list" the Irish sources do not seem to have Gaelicized any of the names, although in the "historic" list, we do find Gaelicized forms in the Irish sources that parallel Brythonic forms in the Scottish sources. This continues to suggest that the "Brythonic" forms are more likely to be original and that Goedelic forms are "translations". (Although by the time of the "historic" list, it might be reasonable to begin finding Irish names being picked up for use by Picts.)

While the earlier lists simply gave single names or genealogies containing only given names, this section begins to contain more constructions that appear to be bynames. Six entries (most of them toward the end) contain a patronym (using "filius" in the Scottish version and "mac" in the Irish -- both, perhaps, translations of whatever was in everyday use). One gives the relationship to a brother, particularly interesting because the brother himself is not listed as having ruled. (In the historic list, there are kings who succeed their brothers are are listed after them as "X frater eius" (X, his brother). Several names appear to have personal bynames. When none of the elements are familiar, as in "Deo ardivois" (which appears in the Irish records as "Deordiuois"), we have to wonder how certain we can be even of the division. (Compare, for example, to "Deocilunon" earlier on the list.) But when one element is known from other names, as in the case of "Gartnaith loc", "Gartnaich diuberr" (probably a scribal error for "Gartnaith"), "Wradech uecla", and "Necton morbet", we can be more confident that we are seeing two separate name elements. Of these, we only have a clue to the meaning of two of them. One of the manuscripts translates "diuperr" as "rich" (Wainwright p.146), and in ColMS, "Necton morbet filius Erop" is referred to in the text subsequently as "Nectonius magnus filius Wirp". The Celtic element "mor" means "large, great", so "magnus" may be only a partial translation of whatever the full meaning of "morbet" was. The name found variously as "Vipoig namet" and "Uipo ignaviet" may also contain a given name - byname combination, but it is hard to tell anything for certain when the sources can't even agree where to break the phrase. If Jackson is correct in connecting the given name with "Vepogenos", then possibly neither is correct.

4. The Historical Kings

The third section of the Pictish Chronicle is the list sometimes referred to as the "historical" kings, because here we have men whose names begin to appear in other chronicles and sources. In the early 6th century, the Irish Dal Riada began settling in Scotland and in the middle of that century they had various interactions with a Pictish king recorded as Brude mac Maelcon, whose death is traditionally placeed in A.D. 583. These events are found in the Irish chronicles of Flann Mainistreach and Tighernac, both dating to the 11th century. (Skene p.cix) So here, again, we have the problem that the events and names are being recorded half a millenium after the fact, perhaps from older written material, perhaps from oral tradition. Chadwick has an extensive discussion of the historic correlations and is, in general, an excellant source of information on the historic aspects of this period.

As usual, the names are adapted to the language in which the record is being made. Just as Brude's name is recorded in the Latin-language Pictish Chronicle as "Bridei filius Mailcon", so he is recorded in the Irish-language version as "Bruide mac Melcon". It should be needless to say that in everyday use we can be as certain that he was not using "mac" as we can be that he was not using "filius". What would he have been using? Either some Pictish word for "son" or a P-Celtic term, either "map" or a more archaic version of the word. One source alone may preserve a Pictish patronym using "map", an 8th century entry in the Annals of Ulster recorded as "Tolarggan Maphan". But, in general, the manuscripts that appear to reproduce the original forms of the names most closely use Latin relationship terms. (For that matter, the Irish chronicle I am using uses "filius" far more often than "mac".) So while we can have a high degree of confidence that the everyday form of the name would _not_ have been any of the recorded versions -- because those languages would not have been in common use among the Picts at that time -- we have extremely little direct evidence for what they _would_ have used.

Continuing with the historical correlations, several names in the middle of the list have connections with Anglo-Saxon royalty of the north. One Talorcan is the son of Enfret (Anglo-Saxon "Eanfrith", a king of Bernicia), the brother of Oswy (a king of Northumberland). Brude, son of Bile, found several names further down the list, is described as "fratruele" to Oswy's son Ecgfrith, suggesting that Brude's mother may have been a close relative of Talorcan. Skene (p.cxxi) suggests a daughter, based on the generations involved, but if the characteristic Pictish succession through the female line was still in effect this would seem unlikely and a niece could be a strong possibility. This list is followed by the introduction of Kenneth MacAlpin, whose reign belongs to the mid-9th century, marking the end of the rule of Pictish kings.

The nature of the name structures has shifted. The vast majority of the names bear a patronym, or some other relationship (one nephew using "nepos", two brothers, although these do not contain the brother's name, but simply refer back to a previous entry, "frater eius"). Four names have no modifier at all, three have what appear to be personal bynames, although none is immediately interpretable.

The list of kings is as follows:

C. Scottish and Pictish Additions to the "Historia Britonum"

There are many other documents containing references to Pictish kings. The majority of them appear to derive from a few early sources, such as the Pictish Chronicle described above, and are generally riddled with scribal errors and corruptions. While there might be some interest in exploring these, they can shed little new light on the topic at hand. A different case can be made for some Pict-related additions to the "Historia Britonum", a legendary chronicle composed around the 7-8th century that gives accounts of the early history of the various races inhabiting Britain. Some of the contents are historical, much is legendary. The best-known early version of this was preserved by Nennius in 858, although there is another from slightly earlier in the same century. The Scottish-related material first appears in a 10th century copy (Harlein MS 3859). (Skene p.xxiiiff) There is some interest in noting the personal names appearing in the Old Welsh king-lists in these manuscripts that correspond to names appearing among the Picts: Eugein, Beli, Elfin, Neithon, perhaps Cinhil. But only four entries specifically identify individuals as Picts, all of them kings. The names don't clearly correspond to individuals in the "historic" king-list (except perhaps for Talorcan), but it should be considered that "king of the Picts" can mean "_a_ king of the Picts" rather than "_the_ king of the Picts". Although the existing copy of this text dates from the 10th century, the nature of the material -- except for very early entries, the material was generally written down contemporary with the event -- and the relatively short lag between the events and the existing copy give us a fairly high confidence level in the forms of the names.

D. OGHAM INSCRIPTIONS

Another potential source of names is an assortment of Pictish stone inscriptions which are, with two or three exceptions, written in the Ogham alphabet developed in Ireland. Based on the particular version of the alphabet used and the style of the accompanying decoration, they can be dated to the 8-9th century - - the very end of the period of identifiable Pictish culture. (Wainwright p.139) Several names found on the stones can be identified with names from manuscript sources.

"Eddarrnonn" (doubling of letters seems common in Pictish inscriptions), or "Idarnoin" in one of the Roman-letter inscriptions, seems to correspond to "Ethernan" or "Ithernan", the first bishop of Rathin and with the entry for "Itharnan" in the Ulster Annals (669). It does not appear to be a Celtic name.

"Drosten" appears in one of the Roman-letter inscriptions.

"Uoret", also from a Roman-letter inscription, corresponds with the Old Breton name "Uuoret" and perhaps with the Pictish king listed as "Uurad" in manuscript.

The Gaelic "Forcus" also appears in a Roman-letter inscription, cognate with "Uurguist" appearing in manuscript.

"Nehhton" appears in Ogham, corresponding to the "Nechton" of the manuscripts, and is almost certainly Celtic.

Other than these names, the only identifiable elements in the inscriptions are the word "crroscc" (from the Gaelic for "cross"), and a number of instances of Gaelic "maqq" or "meqq", the usual Ogham form of "mac" (son) and possibly its genitive "meic".

The unintelligibility of the remainder of the inscriptions -- even when the legibility is perfect -- lends a great deal of support to the thesis that some non-Celtic Pictish language was still in common use at this point. (This makes it all the more maddening that it could survive so late, and yet not be recorded in sufficient quantity that we could know anything about it.) The use of Gaelic "maqq" in these inscriptions does not contradict that theory any more than the use of Latin "filius" in, for example, English documentary forms denies the existence and use of the English language. (Jackson suggests another possibility -- that the Pictish culture may originally have been so lacking in the concept of fatherhood that they had to borrow a foreign word for "son".)

E. MISC. ITEMS

Bede (in the early 8th century) specifically mentions that modern "Kinneil" at the eastern end of the Antonine Wall was called "Peanfahel" in the Pictish langauge, although linguistically it appears to be a blend of P-Celtic "pen" (head, end) and Q-Celtic "fal" (wall). (Wainwright p.143)

Adamnan (in his 7th century biography of the 6th century St. Columba) mentions the names of several people who were, or may have been, Picts. "Emcat" or "Emchat" appears in Pictish country, although he is not identified as a Pict. The name is Celtic, appearing as Gaulish "Ambicatos" and Irish "Imchath". The son of this man is named "Uirolec", which again appears to be Celtic, or at least have a Celtic protheme. "Broichan" is the name of a Pictish king's "druid", and would be cognate with Irish "Froichan", although a more expected Pictish form might be *Uroican. "Brude" or "Bruide" is found in Adamnan as well as in the annals and is discussed in more detail above. Another Celtic name "Artbranan" probably belongs to a Pict. One man specifically identified as a Pict, although living in Ireland, is given the name "Iogenan", which elsewhere is used to Gaelicize the Pictish name "Uuen" (of P-Celtic origin). (Wainwright pp.142-3)

The Annals of Ulster include the 8th century "Tolarggan Maphan", identified as a Pict, which appears to be "Talorcan map ", a rare example where a P-Celtic patronymic construction is found. It also mentions the king "Naiton" which would appear to be the same as "Nechtan" etc. Jackson interprets the chronology of this name as Nechton > Nehton > *Neiton (Naiton). (Wainwright p.145)


Introduction
The Material
Analyzing and Using the Data
Bibliography
Index of Name Elements


Layout, editting, and publishing by Arval Benicoeur.