This volume directly confronts the challenge of reconstructing the dynastic history of early medieval Wales and Ireland on the basis of the extant annals, saints' lives, and king lists, the last of which have been particularly prone to manipulation for political purposes and are thus highly problematic as evidence for the periods they purport to describe. In addition, anthropological studies of modern oral societies have revealed the multiplicity of ways in which traditional accounts of the past, especially genealogies, are routinely re-construed through time to reflect changing ethnic and dynastic relations, or to offer some deeper explanations of the way things came to be as they are. Genealogy was thus one of the key mental constructs through which rulers in the early Middle Ages situated themselves in time and place, and imagined their own social and political identities. Patrick Wormald described the surviving Anglo-Saxon royal pedigrees as a form of "kingly literature" (1986) and the early Celtic kingdoms produced comparable official fictions of their own. Thornton recognizes how difficult it is to use such sources to write the actual history of these houses, without following a counsel of despair in which we simply rule out the royal genealogies altogether as a valid source of historical information.
He argues instead that the king lists can still be used as a highly revealing source of historical inference regarding the periods in which they were composed, and even which they pretend to report, if they are analyzed in a disciplined and systematic way. They can provide "indirect testimony" to the past. He thus proposes a two-fold approach to the records of the early ruling families of Ireland and Wales. The first is to undertake a qualitative assessment of the kinds of bias that may be involved in the construction of any particular genealogy, in response to several questions: (1) For whom and for what purpose was this source first compiled? (2) Why and in what way was it subsequently preserved, transmitted, manipulated, or reinterpreted? (3) How do our answers to these first two questions affect the value of this source as a witness to actual historical persons and events in the times and places they were supposed to have occurred?
Thornton then surveys the kinds of primary source material that must be subjected to this qualitative analysis with regard to two separate questions: (1) the history (and prehistory) of the extant texts as written documents and (2) their developing function through various stages of that history as revealed by the answers to the questions asked above. He then reviews the different types of surviving record for the two countries. Among annals and chronicles, he lists for Wales the Annales Cambriae (c. 1100) and vernacular translations of a lost Latin continuation of Geoffrey of Monmouth's Historia Regum Britanniae (1138); for Ireland, the Annals of Inisfallen, Tigernach, Ulster, Clonmacnoise, etc. In Wales the first extant genealogies appear in the Historia Brittonum (c. 830)--those of Powys, Buellt, and Gwerthrynion--and in the pedigree of Cyngen ap Cadell king of Powys (d. 854) on the Pillar of Eliseg near Llangollen; the earliest Irish genealogies can be found in twelfth-century MSS which preserve material from the eighth century, including some genealogical poetry as old perhaps as the seventh. Both countries also preserve what Thornton calls pseudo-genealogies, by which he means regnal lists rather than genealogies proper, and various origin and explanatory legends. The earliest such for Ireland is the Baile Chuind "Vision of Conn" of the early seventh century, listing the holders of the high kingship of Tara, which was subsequently reworked and expanded over the centuries, and the etiological narratives which are part of senchas "historical learning." And finally Thornton surveys various other kinds of narrative, hagiographical, cartulary, and foreign sources (available in England and elsewhere) to supplement our study of the dynastic history of Ireland and Wales.
This collection of sources makes possible the second step of Thornton's general procedure, one which is more quantitative, adapting the practice of "nominal record linkage" used by modern demographers and psephologists as described by I. Winchester (1970 and subsequently), whereby single individuals are traced by name through all extant records to produce a dossier of potential identifications each of which can be evaluated according to a numerical scheme to assess the relative strength or weakness of their possible linkage. For instance, when comparing similar names in two separate genealogies, Thornton assigns a score of 2 for each direct "match" in four different categories: personal name, patronymic or nickname, coordination in time, coordination in place. He gives a score of 1 for ambiguous information or its absence in one of these four categories in one record, and a score of 0 for a contradiction or "non-match" in any of the four categories in both records. The total possible score is thus 8, which means a very strong positive identification or probability that the same individual is referred to in both records. A score of 4 or less, Thornton suggests, should prompt rejection of the apparent linkage: 5-6 yields a "possible" connection; 7 one that is "preferable" to its rejection; and 8, as before, suggests a very good probability that both sources refer to the same ruler. This system works especially well for Irish sources where the information supplied is usually somewhat fuller than in the Welsh, which tend not to include a patronymic or kingdom, so that the linkages in Welsh records will often never be capable of a reliability rating higher than 7. Even so, this system of quantitative analysis yields a set of possible references to single named rulers of varying but ascertainable reliability that can then be studied in concert to draw larger inferences. The genealogies, in concert with the chronicles, saints' lives and other sources, can thus be coaxed into providing valuable suggestions about the way the history of the dynasty was conceived when they were compiled and sometimes, though more generally and often only negatively, about the actual history of that dynasty in the earlier period they claim to describe. At the very least we can confirm in some cases that certain persons and events as recounted in these sources must certainly be political fictions not worthy of further consideration as historical facts.
Thornton demonstrates the fruits of this double approach in detailed studies of the origins and development of three individual royal families: (1) the second dynasty of Gwynedd in North Wales, putatively founded by Rhodri Mawr; (2) the Deisi Muman of Munster in Ireland, which was supposedly also responsible for the establishment of the ruling house of Dyfed in Wales; and (3) the Conaille Muirtheimne of County Louth in Ireland. Thornton hopes his analysis of the records relating to these royal families will not only demonstrate the usefulness of his method, but also contribute substantially to our historical knowledge of these three kingdoms and their ruling houses.
In "Fratricide, Filiacentricity, and the Rise of the Second Dynasty of Gwynedd," Thornton studies a family that emerged as rulers in North Wales in the earlier ninth century extending its power over most of the rest of the country by the middle of the tenth. He deconstructs several of the foundation legends associated with this house, including (1) the supposed fraternal relationship between Hywel and Cynan whose conflict in the 810s made the first dynasty vulnerable to a coup by Merfyn Frych; (2) the claimed existence of eight sons of Rhodri; and (3) the association of this dynastic progenitor with the northern Brittonic-speaking kingdom of Manaw in southern Scotland. Thornton concludes that the genealogical records of the second dynasty of Gwynedd are the product of both contemporary manipulation and later antiquarian speculation. For later periods, it is possible to reconstruct some dynastic sequences with reasonable confidence, but further confidence in these sources without independent verification is not possible. Thornton finds this result rather disappointing. His scrutiny of the earlier Welsh genealogy of the dynasty of Rhodri Mawr has rendered the period it describes more, rather than less, obscure.
His results are more positive when he examines "The Deisi Muman in Munster and Dyfed," particularly as described in the eighth-century Indarba na nDeisi "Expulsion of the Deisi," a term for "subject or client tribes," which contains a fictionalized account of the migration of one of these groups from County Meath to Munster, where they supposedly first founded their kingdom in the fifth century. Thornton finds the political purpose of this text to be quite demonstrable: it is a fiction invented to explain why the kingship is controlled by segments of the clans Ui Eogain and Ui Rossa to the exclusion of claimants from the collateral Ui Oengusa. He notes an interesting fact that the related saints' lives include reference to many kings of the Deisi Muman not found in the genealogies, suggesting either that the hagiographers invented them for their own narrative purposes or that they represent actual kings who had lost their political usefulness to later royal genealogists and were thus only remembered incidentally by the hagiographers. Thornton then reviews the good amount of information that can still be gleaned about the early rulers of the Deisi Muman during the seventh and eighth centuries before the death of Dunchad mac Eogain in 764, after which the kingship was "concentrated in the hands of his descendants" (168), with much suppression of complex genealogical material into "a single (and somewhat confused) line of descent" recorded in a unique source (179). This middle period of the history of the dynasty is thus very obscure until c. 900, after which more generous sources from rival factions of the Ui Fhaelin and Ui Bricc allow Thornton to complete his account of Deisi Muman more confidently through the eleventh century.
Thornton also examines in this chapter a question of great interest to Welsh scholars, the authenticity of the legend, preserved in the earliest version of the Indarba, that the royal dynasty of Dyfed was founded by a migrant branch of the Deisi Muman. The Irish origin of this Welsh house has become almost axiomatic among students of the period, but Thornton believes it cannot be true. He offers an explanation as to how the legend was intended to function politically. 1. The southern Welsh rulers of Dyfed wished to distinguish themselves from their competitors to the north in Gwynedd, who claimed (as we have seen) that their ancestors had arrived from the further north of Britain to expel the Irish from Wales: "Northern British-Irish opposition in the sub-Roman period may have been used as a means of mirroring the later opposition between northern and southern Welsh kingdoms" (154). 2. We know that St. David's in South Wales had contact with southern Ireland, where the figure of Eochaid Allmuir appeared in an earlier reformulation of the Deisi Muman genealogies as a residual ancestor without descendants: this circumstance may have been rationalized by the assumption that he was one of the migrants to Wales in an earlier period and "[c]ontact with Dyfed may have supplied a suitable pedigree onto which Eochaid could be conveniently grafted" (155). 3. This enhanced genealogy was borrowed back by Dyfed so that it could claim a distinguished foreign ancestry from Ireland to promote their claims to national authority in Wales in answer to the boast of the second dynasty of Gwynedd of their distinguished foreign origin in the north of Britain.
As his last test case, Thornton analyzes the "schizophrenic" set of pedigrees pertaining to the "Conaille Muirtheimne as Buffer-Kingdom," which dynasty first appears in the late seventh century and disappears at the end of the eleventh. He argues that this family was established in County Louth as a client of the Ui Neill of Ulster to help contain other rival clans in that country, specifically the Ulaid to the northeast: "Over time, however, political expediency (perhaps, in part, due to the encroachment of the Airgialla) forced the Conaille into closer relations with, and eventually dependency upon, the Ulaid," whom they were originally intended to check (208). The extant genealogies reflect this shifting political orientation, revealing several contradictory attempts to associate the prehistory of their house with one or the other of their Ulaid or their Ui Neill overlords. Thornton attempts to reconstruct the dynastic history of this kingdom with his usual careful assessment of the probability of the various identifications offered and concludes that despite their own serious attempts to integrate themselves more deeply with rival Ulster groups, the Conaille Muirtheimne never really became part of the Ulster kingdom, whether within the Ulaid or the Ui Neill spheres of influence.
Thornton's examination of the genealogical records of these three dynasties is extremely detailed, not to say somewhat tedious and hypothetical, at least for Wales, so that not all readers will wish to follow him through the intricacies of sorting through the evidence in the painstaking way he pursues. Most will be content to take Thornton's general approach to early royal genealogies as a balanced and sensible way of using them as historical documents. He concludes that without independent verification, genealogical lore from these kinds of record must always be deemed potentially unreliable. On the other hand, incidental information which does not seem to serve any obvious ideological purpose should be tentatively respected as "potentially reliable" (228), since its inclusion may not have been motivated by any other factor than that it represents the compiler's understanding of the facts. Such sources thus present at least a prima facie case, which methodical analysis can assess as certainly fabricated, rather improbable, to be preferred among other possibilities, quite probable, or in a few rare instances, almost certain to be the case. Even so, Thornton warns severely against taking any of these records at face value for the periods of history they pretend to describe, a failing he finds particularly unfortunate with regard to the unhistorical legend of the Irish origins claimed for the royal house of Dyfed.
The author ends his study with a differential assessment of our capacity to trace the history of the early dynasties of Ireland and Wales. Irish dynastic history "can still be reconstructed using the surviving genealogical and pseudo-genealogical sources, though the resulting origin may be very different from that formally outlined in these documents" (208). The genealogies of the Irish dynasties have been preserved in much greater detail about collateral and rival lines than the Welsh pedigrees, which are often factitiously resolved into single patrilines to bolster the claims of a particular ruler promoted therein. Thornton notes, however, that the Welsh annals and dynastic records are available in much better editions, especially Peter Bartrum's Early Welsh Genealogical Tracts (1966), while "a large part of the corpus of medieval Irish genealogies has not been published and is accessible, at best, in facsimile form. Similarly, though less drastically perhaps, many chronicles can be consulted only in out-of-date and often difficult editions" (229-30). Efficient further analysis of these records according to the methods Thornton proposes thus awaits the scholarly production of Irish genealogies in more conveniently usable forms.