A Consideration of Pictish Names

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

III. Analyzing and Using the Data

A. Ethnographic Considerations in Analyzing the Name Data

There are a wide variety of "origin legends" for the Picts, most of which seem to be motivated by a desire to connect them with every known appearance of the term "Picti" or "Pictones" (never mind that these are Latin terms) or with the custom of painting or tattooing the body. But one aspect to these origin legends that holds steady is a claim that Picts traced succession through the mother. The king-lists bear this out, showing frequent occasions where brothers would succeed each other, but no instance of a son succeeding his father. The legendary histories come up with a wide variety of "explanations" for this, but probably only the observation of the phenomenon itself is dependable. (Skene p.ci)

In at least two cases, we have records showing Picts (or probable Picts) recorded using the relationship "nepos". While this originally meant "grandson", it early acquired a secondary meaning of "nephew". We find it in the 3rd century Colchester inscription "Lossio Veda ... nepos Uepogeni" and in an entry in the "historic" king-list "Nectu nepos Uerd". If inheritance were being traced through female lines, then the salient male-male family relationship might be expected to be that of uncle and nephew (sister's son).

Another set of related observations from the king-lists is made by Skene (p.ci): "First, the names of the fathers and of the sons are quite different. In no case does the name borne by any of the sons appear among the names of the fathers, nor, conversely, is there an instance of the father's name appearing among the sons. Second, the names of the sons consist of a few Pictish names borne by sons of different fathers. There are 6 Drusts, 5 Talorgs, 3 Nectans, 2 Galans, 6 Gartnaidhs, 4 Brudes. In no case does the name of a father occur twice in the list of fathers. Third, in the list there are two cases of sons bearing Pictish names whose fathers are known to have been strangers, and these are the only fathers of whom we have any account. They are: 1. Talorg Mac Ainfrit. His father was undoubtedly Ainfrit, son of Aethelfrit, king of Northumbria, who took refuge among the Picts, and afterwards became king of Northumbria. 2. Brude Mac Bile. His father was a Welshman, king of the Strathclyde Britons. In an old poem, Brude Mac Bile is called son of the king of Ailcluaide, i.e., Dumbarton; and when, by the battle of Dunnichen, he became king of the Picts, another old poem says, 'today Brude fights a battle about the land of his grandfather.'" Skene goes on to discuss the possible origin of these facts in a system of exogamy with matrilocality. In strict point of fact, Skene has missed the appearance of "Unuist", "Talorgen", and "Constantin" in both father and son positions toward the end of the "historical" king- list. But his observation is generally sound: the given names of the Pictish kings center around a group of "popular" items, most of which are strongly connected with Pictish culture in other sources (Pictish Oghams, etc.). The patronyms are a less coherent group, and while a pattern of exogamy in Pictish royal families might frequently result in non-Pictish patronyms, it would not necessarily always do so. While we cannot assume that the patronyms in these lists are Pictish, neither can we automatically assume that they are non-Pictish.

Another observation from this material is that, whatever the original Pictish attitude towards fatherhood and its acknowledgement may have been, by historic times (i.e., around the 6th century) it appears that kings' fathers were known and that this knowledge was incorporated in many of their recorded names. We should not be too quick to take this at face value. It is certainly possibly that some of these "patronyms" originally referred to uncles, not fathers (as with the entry using "nepos") but that these were reinterpreted by later generations for whom patronyms were much more the rule. It is also possible that the names refer accurately to the kings' fathers, but that the insistance on using a patronym in the records, rather than some other relationship, is dictated by later (or foreign) sensibilities. It is also perfectly possible that while the method of inheritance remained the same, the Picts themselves had adopted Celtic ideas of naming to the extent of perfering patronyms to some system that made more functional sense in their society.

It is important not to fix on any of these as the "true" situation without further evidence; I can do no more than sketch out some of the possible scenarios. This may be fine for historic analysis, but for re-creation, some sort of decision has to be made. In the following sections, I present advice and guidelines on constructing plausible Pictish names, both for the Roman era (ca. 1-4th centuries) and for the medieval era (ca. 6-9th centuries). Generally, I will present several options that seem reasonable, based on the data. These may not exhaust all the plausible constructions, nor may some of them be correct (did we but know all the truth). Thus they should be taken, not as any sort of definitive answer, but simply as a bit of educated guesswork.

B. The Roman Era (and Latinate Constructions in General)

Since the question that prompted this research was on the question of 1st century Pictish names, I have given a disproportionate amount of consideration to this era. My basic premises are these:

- While it is reasonable (and perhaps necessary) to assume a Latin-language context for written names of this era, for the 1st century, it is not reasonable to assume that Latin-origin name elements would have entered the available pool.

- We do not know anything useful about non-Celtic Pictish grammar, and so the best we can do in constructing early Pictish names is to give them either a Latin or Celtic context in terms of "function words" (such as relationship terms) and inflectional endings. In the case of a Latin context, we have clear exemplars of how native British names were treated. In the case of a Celtic context, there will still be a certain amount of guesswork in the choice of function-words and word order.

- The evidence shows, in general, that Celtic-origin names were in widespread use by the Picts, and the earliest recorded names of probable Picts are of Celtic origin; therefore, while it is interesting to note which names cannot be clearly identified as Celtic in origin, it is not appropriate to suggest that Celtic- origin names would be incorrect for Picts.

- The best interpretation of the evidence is that the Celtic influences on Pictish langauge and names were of the P-Celtic (Brythonic) variety before the 6th century immigration of the Dal Riada, therefore "Gaelicized" forms of given names and Q-Celtic constructions (such as the use of "mac") should not be considered appropriate for the pre-6th century period.

1. Given Names

For given names, we have only four names from the Roman period that are clearly located in the Pictish sphere, and those are clearly Celtic in origin.

In addition, we can take some of the Celtic names, at least, from the medieval king-lists and (with the help of Jackson's analysis of the sound changes) make a fair reconstruction of a possible Roman-era form. These are found in section III.C.1. following.

In addition to the above names that are at least connected in some fashion with Pictish people or lands, we can also expand our consideration to other Celtic names found in Britain. This expansion ignores the fact that there were different groups of Celts in Britain -- e.g., the more recent Belgic people, largely in the south and east, versus the older non-Belgic inhabitants -- which might well have significant differences in their name- stocks. Such a consideration is beyond the scope of this paper. I simply present here a list of British given names (a few may instead be Latin -- it's sometimes hard to tell) from Roman-era inscriptions taken from my article "The First Thousand Years of British Names". (These are from Burn and Birley and cover the period up to the 4th century. The two sometimes conflict with the assignment of language to a name, such as for "Vindex", and I have tended to be liberal in this list. In some cases, my source has reconstructed the British ending "-os", in others it is left in the Latinized form "-us". In some cases, the nominative form has been reconstructed from whatever was found in the original inscription, usually a genitive. Women's names are indicated by (f) after the entry.)

As a quick-and-dirty approximation to genitive (posessive) forms, replace as follows:


2. Locative/Tribal Bynames

The tribal and place names of the Ptolemy map can be converted into the sorts of forms found in Latin locative adjectives. I have tried to present both a survey of the constructions, and tables of grammatical forms for the Ptolomeic names. The constructions are taken from Cassell's Latin Dictionary and from Burn, which contains inscriptions found in, or relating to, Britain. The grammatical forms are my own, and I will freely admit that my comfort level with Latin idioms is not the best. While several options are given for each name, in actual practise we often find that only one particular form would be used in a particular case (e.g., always "Romanus", never "Romensis"), while in other cases, various constructions can be found (e.g., Helveticus versus Helvetius, Brittanicus versus Brittanus). These tables should be taken as a general outline only, and suggested corrections or modifications are welcomed.

When tribal names are used for personal bynames (or as nouns indicating a person of that tribe, which is functionally identical) this can be done in a number of ways. In some cases, the tribal name is taken as a simple plural of the singular noun referring to one member. Thus we find "Gallus" and "Galla" respectively referring to male and female Gauls, for which the tribal name is "Galli". But we also find the tribal name converted to an adjective and thence to a noun, in the cases I have seen, using the suffix "-cus", e.g. Helvetii > Helveticus, Germani > Germanicus. Another form found in inscriptions uses "natio" (tribe) either in the genitive (nationis) or what appears to be the ablative (natione) followed by the tribal name in some form. Unfortunately, the examples I can find of this are either abbreviated, corrupted, or simply inexplicable. "Natione Belga" could be an ablative of origin using an adjectival form "Belgus", i.e. "from the Belgian nation", perhaps; similarly "Natione Cat[u]vallauna". But "nationis Brigans" using the genitive of "natio" (of the tribe) looks like it uses a nominative singular, back-formed from "Brigantes" (treating it similarly to "pons/pontis"), where we might expect the genitive plural ("nationis Brigantum"), which makes no grammatical sense whatsoever. Further research into this construction should be done before drawing conclusions.

The following table gives forms based on the tribal names, showing a hypothetical nominative singular (both masculine and feminine), nominative plural (the form provided by the map), genitive plural, and adjectival form using "-cus" (the masculine, for which the corresponding feminine will be "-ca").

Personal bynames or descriptions taken from town names can also take several forms. Inscriptions are notorious for cryptic abbreviations, but it appears that several of them simply use the place name in the nominative to signify origin. (E.g., "C. Lovesius Papir. Cadarus Emerita", Caius Lovesius Cadarus of the tribe of Papiria, from Emerita.) Another construction uses the genitive of the place name, i.e., "belonging to ", as in the formula "G. Valerius G. F. Galeria Victor Lugduni" (Gaius Valerius Victor, son of Gaius, of the Galerian voting-tribe, of Lugdunum). A far more common method, however, is to form an adjective from the place name using "-ensis" or more rarely "- anus", the latter being seen in "Romanus", i.e. "a Roman (man)" and in a variant in "Barates Palmyenus" (Barates, the Palmyrian), the former in examples like "L. Vitellius ... Tancinus Cives Hisp. Caurie[n]sis" (Lucis Vitellius Tancinus, citizen of Spain, a Caurian, i.e. from Caurium).

The following table contains the presumed town names from the Ptolemeic map showing the nominative singular (the given form), the genitive singular, and adjectival forms in "-anus" and "- ensis". The feminine version of "-anus" will be "-ana". "-ensis" is both masculine and feminine. To somewhat oversimplify, roots ending in "-is" or "-es" can take adjectives in "-ianus" or "- inus", which I have indicated by "-i(a)nus", and probably in "- iensis", although I haven't been able to track down specific exemplars for this. I have eliminated those town names that are Latin in origin.

Dumna Dumnae Dumnanus Dumnensis
Scetis Scetis Sceti(a)nus Scetiensis
Orcades Orcad(i)um Orcadi(a)nus Orcadiensis
Malaius Malaii Malaianus Malaiensis
Tarvedum Tarvedi Tarvedanus Tarvedensis
Virvedrum Virvedri Virvedranus Virvedrensis
Verubium Verubii Verubianus Verubiensis
Carnonacae Carnonacarum Carnonacanus Carnonacensis
Creones Creon(i)um Creoni(a)nus Creoniensis
Tuesis Tuesis Tuesi(a)nus Tuesiensis
Bannatia Bannatiae Bannatianus Bannatiensis
Tamia Tamiae Tamianus Tamiensis
Devana Devanae Devananus Devanensis
Epidium Epidii Epidianus Epidiensis
Alauna Alaunae Alaunanus Alaunensis
Orrea Orreae Orreanus ?
3. Descriptive Bynames

In addition to these locative/national byname possibilities, we have one example of a probably-descriptive byname borne by a "Caledonian" in a Roman-era inscription. Jackson interprets "Ueda" as perhaps meaning "knowing", but not much weight should be put on this. His interpretation is based on an assumption that the word is Celtic in origin.

4. Relational Bynames

The same inscription from which our sole descriptive byname is taken provides a formula that may well be characteristically Pictish, naming as it does a relationship to an uncle rather than a father. Later names clearly note that, although inheritance was said to pass through the mother, scribes were recording names with patronyms. When the shift occured that changed this convention (or even whether patronyms naming the father might have existed in parallel with our early "nepos") is impossible to know. It is certainly possible to suggest for this era:


It may also be valid to suggest:


If descent and inheritance are being traced through the mother, it may also be valid to use the latter construction with a mother's name. The Latin feminine forms of the above would use "neptis" and "filia" -- although given the complete lack of Pictish feminine names, there hardly seems any point in mentioning it.

C. The "Medieval" Era

1. Given Names

The following table shows the given names found in the post- Classical records that have the strongest presumption of being "Pictish", i.e., eliminating names found only in patronyms (which may belong to non-Pictish fathers) or those which appear to be only legendary. The actual forms taken from the records can be found in the index at the end of this paper. In this table, I present "standardized" forms representing a reconstructed "Classical" form (contemporaneous with the Roman-era forms), an "early-medieval" form, and a "later-medieval" form, intended to represent roughly 6th century and 8th century forms. I have only given reconstructions where I have some level of confidence in the results. Many of the "Classical" reconstructions are from Jackson, rather than being my own. Reconstructed forms are marked with an asterisk. Ideally, actual cited forms from the data should be used in preference to these, but the table of forms can be used to guide the choice of spellings compatible with earlier or later periods.

Before giving the alphabetic table, it's worth noting the most popular names. This "popularity" list includes the patronyms.


Pictish Given Names From Medieval Sources
Classical6th C.8th C.
Given Names Appearing in Patronyms
The given names appearing only in patronyms I have treated as above. Keep in mind that the fathers of Pictish kings may not themselves have been Pictish. In one case, the name is clearly of Anglo-Saxon origin (Enfret).

Classical6th C.8th C.

2. Bynames

It is hard to know how to present the bynames appearing in this material. The vast majority are of unknown meaning -- and unknown levels of scribal corruption.

Diuberr             "rich"
Erilich             or possibly "Erilith", unknown
Gaed brechach       unknown
Ini                 unknown, possibly scribal error
Loc                 unknown
Morbet              "mor" seems to be "great, large" but the rest
                    is unknown
Olgudach            unknown, except that "-ach" is often found
                    forming Gaelic adjectives
Uecla               or possibly "Uetla", unknown
Gurthinmoch         unknown

3. Construction

The basic "historic era" name constructions we see are:

"Mac" is the Gaelic patronymic marker and would have come into use with the gradual Gaelicization of Pictish culture. It almost certainly would not have been in common use in the 6th century, and almost certainly _would_ have been in common use towards the 8th century. Patterns of use between those periods are anybody's guess. "Map" is the Brythonic equivalent to "mac" and may have been in use during the 6th and 7th centuries and earlier. There is only one questionable example of it. "Filius", "nepos", and "frater" are all Latin terms and might have been used in written records, but most likely not in common speech. The Old Irish equivalent (and cognate) of nepos is "n{i'}a", and this shows up in an Irish Ogham inscription in the genitive "niottas" (which isn't a form that would normally appear in a name, but it shows that it was used). The Old Welsh form would be "nei", and a reconstructed Brythonic form (compatible with the Classical name forms) would be "*neis".

Another plausible type of byname would be the locative type discussed in the Classical section above. In the absense of vernacular examples, the safest suggestion would be to use a Latinized form, with the "-ensis" suffix added to a Latinized form of the place name. This construction can be found in medieval Latin documents. It is beyond the scope of this paper, however, to list Latinized versions of early Scottish place names.

IV. Bibliography

Bartrum, P. C.. Early Welsh Genealogical Tracts. Cardiff: University of Wales Press, 1966.

Birley, A. R. Life in Roman Britain. London, 1964.

Burn, A.R. The Romans in Britain: an Anthology of Inscriptions. Columbia: University of South Carolina Press, 1969.

Chadwick, H.M. Early Scotland. Cambridge: Cambridge University Press, 1949.

Henderson, Isabel. The Picts. London: Thames and Hudson, 1967.

Ritchie, Anna. Picts. Edinburgh: HMSO, 1989.

Skene, William F. Chronicles of the Picts, Chronicles of the Scots, and other Early Memorials of Scottish History. Edinburgh: H.M. General Register House, 1867.

Wainwright, F. T. ed. The Problem of the Picts. Edinburgh: Thomas Nelson and Sons, Ltd., 1955. (especially the chapter by K.H. Jackson "The Pictish Language")

The Material
Analyzing and Using the Data
Index of Name Elements

Layout, editting, and publishing by Arval Benicoeur.