2: The River Severn/Hafren and Caerleon/Caerlleon

Once again, Geoffrey of Monmouth is blamed for linking Habren with Sabrina (HRB ii.5), despite the etymological connection between Brittonic *Σabrīnā and Welsh Hafren by way of Old and Middle Welsh Habren; the Σ is a highly aspirated sound, heard by Latin speakers as an S, in which form it passed into Old English. The identity of the two names (linguistics aside) is shown in the Historia Brittonum Chapter 68, which describes duorig habren, id est duo reges sabrinae (Dau Ri Hafren, that is the ‘two kings of the Severn’), the phenomenon of the Severn Bore.

However, Blake and Lloyd use the name Caer Loyw in Culhwch a Olwen to suggest an alternative location for the Hafren. According to the Brut, Caer Loyw was founded by a Roman emperor named Gloyw, whom they identify with Claudius. This is a garbled version of the story in Geoffrey of Monmouth (HRB iv.15) whereby Kairglou is founded either by Claudius following the marriage of his daughter Genuissa or by Dux Gloius, whom Claudius fathered there. However, they then make a connection between Arvirargus, husband of Genuissa in Geoffrey, his son Marius (HRB iv.17) and a ‘tradition’ that he built the walls of Chester. This they then take as evidence that Caer Loyw was Chester. This is backed up by a line from Gutyn Owain, writing c 1470, that at the battle of Chester (which they confusingly and uniquely refer to as the Battle of Bangor), Brochwel was forced to retreat through the River Afren, which thay take to be the same as Hafren. However, the River Dee was occasionally known as Aerfen in Middle Welsh, and it seems either that Gutyn Owain or his editor was guilty of a spelling error.

In fact, Caer Loyw is the Middle Welsh form of a name found in Old Welsh sources as Cair Gloiu (for instance, Historia Brittonum Chapter 66a), showing the loss of initial G- through the perfectly normal process of mutation, a regular feature of Middle and Modern Welsh. The Historia Brittonum Chapter 49 contains the staement that gloui… aedificauit urbem magnam super ripam fluminis sabrinae, quae uocatur brittannico sermone cair gloiu, saxonice autem gloecester (‘Glous… built the great city on the bank of the river Sabrina which is called in British speech Cair Gloiu, but Gloecester in English). Here we have an identification at the start of the ninth century of Gloucester with Caer Loyw. This is also the source of Geoffrey of Monmouth’s statement about the origin of his Kairglou and the Cair Loyw of the Brut. The name derives from the Romano-British Gleuum, the Colonia (earlier legionary fortress) at Gloucester. This in turn derives from a Brittonic *glēuo-, meaning ‘bright’; the word survives with the same meaning as Welsh gloyw/gloew.

Even more damning for Blake and Lloyd’s identification of Hafren with the River Dee is their identification of Caerlleon in the Brut with Chester. Although Chester was certainly known by this name in Middle Welsh, the Brut specifies the name as Caerlleon ar Wysg. Wysg is the Middle Welsh form of the name of the River Usk in Gwent, on which the south Welsh Caerleon stands, and it derives from Brittonic *Īscā, found in Roman sources as the name of Caerleon, the legionary fortress that stands on its banks. They find the statement that Caerlleon ar Wysg was an archbishopric a problem for interpreting it with Caerleon on Usk (which they call the ‘Roman town of Isca’ – it was in fact a legionary fortress), as there is no record of there having ever been an archbishopric at Caerleon. They go on to state that Chester was made the centre of a palatine earldom in 1071 (current opinion is that the palatinate status of the county was something that accrued over many years) and that it has been the seat of a bishopric ever since. The bishopric of Chester dates only from August 1541; a short-lived medieval bishopric was created in 1075, when the former Bishopric of Lichfield was transferred there, but this was moved to Coventry by Robert de Limesi in 1102. This is irrelvant to their argument.

They also draw in a large number of medieval (and later!) texts to establish that Caerlleon ar Wysg was Chester. These texts show a confusion between the two sites that is not found in the earlier sources. Moreover, they use an early eighteenth-century text for linguistic analysis of Wysg to suggest that the name Wysg could also be applied to the Dee. The word Wysg (along with other river names such as Axe, Exe, Esk and so on) derives from Brittonic *ēscā, which seems to have arisen through the influence of Latin on an earlier *ĭscā, meaning ‘water’. The eighteenth-century writer quoted by Blake and Lloyd, Theophilus Evans, makes the equation between Wysc and a word ‘visc’ used ‘by the Gwyddel of Ireland… for Dwfr’; in fact, the word occurs as uisg in Gaelic to mean ‘water’ and is found in an early Irish glossary in the form esc. What this shows is that the Brittonic word *ĭscā, ‘water’, has a cognate in Irish; so does dwfr, which occurs in Old Irish as dobur and modern Gaelic as dobhar. In other words, both languages have more than one word for water. This is not surprising. However, the evidence of placenames suggests that those rivers regarded as being *ĭscās were not interchangeable with those regarded as *dŭbrās. The former have names that survive as various Axes, Exes, Usks and so on; the latter include the River Dee (Welsh Dyfrdwy, literally ‘waters of Dee’, *dŭbrās dēuās).

Finally, they make the unsupported claim that ‘Chester was… the major city in the Kingdom of Gwynedd’. In fact, the kingdom of Gwynedd never controlled territory farther east than the Vale of Clwyd until the medieval wars between England and Gwynedd. Chester appears to have been part of one of the territories that became known as Powys during the early Middle Ages. Indeed, the presence of Brochwel, a member of the royal dynasty of Powys, at the Battle of Chester (whom they refer to when dicussing the ‘Battle of Bangor’) confirms that Chester was not part of Gwynedd.