The Latin origin of the word 'river' was riparia, meaning the banks of the water.

In Early French it became rivere, and represented both the banks and the water between them. That became the source of the modern English word, river.

The current Italian descendant of the word still refers to the shore, and the Riviera is a region along the coast of the Mediterranean Sea. The older Latin version has not disappeared and remains mostly in legal terms related to the rights of those who control the banks of the river, called riparian rights.

Closely related is the word 'rival' - those who share a common stream. The original meaning was closer to our present word for companion.

The word 'arrive' also comes from the same root ripera. As a river bank ends when the river reaches its destination, so a journey ends when we arrive. And speaking of words that are derived from other words, derive means, literally, from the stream.


Place-name scholars describe most English river names as ‘Celtic’ or ‘pre-English’, so would appear to be suitable for interpretation in terms of Archaic Gaelic. A closer look at the current approach to English place-names shows that it is based on some rather quirky assumptions which are by no means compatible with the Archaic Gaelic hypothesis.

According to that view, English settlement names, with a few (but significant) exceptions, were created in the post-Roman period by invasive peoples coming into England from northern Germany or from Denmark. River names, however, were passed on by displaced or enslaved native peoples. A similar belief holds in Scotland, where all place-names are attributed to Irish, Welsh, Northumbrian and Norwegian settlers, with virtually no native survival, and in France, where Franks and feudal landowners are implicated. In every case the language of settlement names is taken to be the contemporary written languages of these various peoples.

When one takes a wider view it becomes evident that there is a conflict between the approaches. When current theory suggests that we should find shared features and similar names, we do not. The names attributed to Scandinavian settlers in the north and west of Scotland have little or nothing in common with the names proposed for the Solway, Cumbria, or Yorkshire, or, for that matter, those found in Iceland or Scandinavia.

Where there should be divergence, we find common features. The same basic elements occur in Scotland, England, France, Flanders and further afield in names that allegedly reflect settlement by different linguistic groups.

Settlement names share a great many elements with river names, which suggests that the two classes must be of similar age. A further discrepancy is while the English place-name lexicon may sometimes resemble Anglo-Saxon it does not do so consistently, and many of the interpretations in terms of Anglo-Saxon are, to say the least, unconvincing. All of this suggests an older common origin for British place-names and a common origin for river names and settlement names.

Place-name theory is showing its age. It is increasingly adrift from the findings of archaeology and early history. Archaeologists no longer believe in the kind of population replacement proposed for England and Scotland, and it is also declining in popularity with historians, who are waking up to the very dubious historical value of their early sources. Incomers can certainly be detected in post-Roman England, but they appear on the whole to have been humble folk and at the most optimistic calculation could not have added more than five per cent to the existing population. Whether elite or slaves, there is no evident reason why these immigrants should have altered a settlement pattern which in many places can be shown to have been in place before the Roman occupation. If the settlement pattern survived, it is a safe assumption that settlement names also survived.

We might accept a few signs of Norwegian administration in the former Norwegian colonies and a few new farms carved out of the of northern Europe but not wholesale resettlement. English settlement names certainly look ‘Anglo-Saxon’ rather than ‘Celtic’ and many of them lend themselves to interpretation in terms of Germanic personal names and Anglo-Saxon, though the results are often very unconvincing.

The explanation lies in the history of literacy in England. Those who first transcribed English settlement names and who perpetuated these early written forms through the following centuries were not natives but foreigners whose native language might have been Frankish, Danish or French and who were literate in Anglo-Saxon, French or Latin. They were detached but not neutral. Like the nineteenth-century surveyors who first transcribed many Gaelic place-names in Scotland, they were influenced and restricted by their native languages in their choice of forms and they often ‘recognised’ personal names and other elements of their current vocabulary which did not exist in fact. A Scottish example is Cyderhall (Sutherland), which has recently reverted to an earlier spelling Sythera, suggesting an original outlook post or sith rather than an English apple orchard. However this place in the thirteenth century was ‘Siwardhoch’ or Sigurdhaug ‘the burial place of Sigurd’. A more recent example is the name Thomasgreen, found in the township of Braeleny near Callander, Perthshire. Thomasgreen was once the site of the school was and can be reconstructed as tom a’ sgriobhan, ‘the knoll of the scribbling’.

A further source of bias is that English clerks have always been familiar with the Anglo-Saxon foundation fable. It was created in the ninth century by foreign clerks who used as evidence an excessively naïve interpretation of a handful of English place-names in terms of Germanic personal names. Naïve or not, since then everyone concerned has continued to interpret English settlement names in terms of Germanic personal names or in terms of the obscure written ‘Anglo-Saxon’ used in religious texts and a few secular documents from the eighth century onwards.

However, the use of personal names to identify local farms and hamlets is not very common in the English-speaking world. Those who came to England in the post-Roman period were illiterate and pagan. One might wonder how such transient personal names were perpetuated? And we might also wonder if these illiterate settlers, whose origins are diffuse and obscure, spoke a language which is known only several centuries later in a literate, elite, Christian context? Such questions are not asked. The place-name creed assumes settlement by Saxon invaders and explanation is a secondary matter. There are certainly links between place-names and personal names, but they arise from the fact that European personal names were very often derived from the names of places, as is still often the case, at least in Scotland. But the landscape was named first.

A striking example of the ability of imported clergy to influence or reinterpret local place-names in terms of personal names is found not in Anglo-Saxon England but in the Celtic south-west, where Breton clergy, in a frenzy of pious creativity, converted dozens of place names into those of parish saints, most of whom are known nowhere else in Christendom. They found Sulyen at Luxulyan, Mabat Lavabe, Mawgan at Mawgan, Melanus at Mullion and a great many more. It appears that the Christian forms were largely confined to literary use, kept alive in documents maintained by the Church while the natives continued to use the original names on which such speculation was based. These vernacular names have persisted in many cases. Towednach is now believed to be a broken-down version of the correct form, ‘Saint To-Winnoc, ’ which it barely resembles but is readily explained as ‘gathering fire’, a very common type of settlement name. Towinnoc is also ‘gathering fire’ but with Archaic Gaelic, fionn ‘fire’ instead of aodh ‘fire’. The parish names of Cornwall and Devon show that the influence of literacy on the form and apparent meaning of a place-name can be very powerful. Finally, current studies cannot explain how it is that many settlement names recorded by the Romans and other early writers remain in use. Even York, in its Latin form Eboracum or its Danish form Euruic, remains recognisable. It is surely curious that the settlement names recorded in the Roman period are the only settlement names which are allowed to have survived from before the advent of the Saxons. It is entirely improbable that in order to have remained in use, a pre-English settlement name must have been written down in the Roman period.

When we use Gaelic and Archaic Gaelic to explain English river-names, a very different picture emerges. The names make consistent sense as early names coined by hunters. The key elements are tribal deer forests and fire, either cooking fires indicating settlement or, more probably, beacons used to define territory and organise hunting. Explanations currently offered for English river names include many ‘water’ words such as ‘fresh’, ‘clean’, ‘bright’, ‘running’, ‘dirty’, ‘pure’ and ‘smooth’, but such words, which define transitory states, are without value as placenames. The alternative explanations offered here are of course tentative but make consistent sense within the parameters of a hunting population and relying on what is known of Archaic Gaelic in other contexts.

With the exception of names such as Avon, Don or Esk which mean simply ‘water’ and which are very early indeed, river-names are not a class apart but part and parcel of the basic naming pattern. Almost every river name listed below is matched by a number of settlement names. Kenn is the name of two rivers but is also found in more than thirty settlement names listed by Mills. Laver is the name of a river in Yorkshire but occurs also in Laver (Essex), Laverstock (Wiltshire), Laverstoke (Hampshire), and Laverton (Gloucestershire, North Yorkshire, Somerset). Sid is a river in Devon but also occurs in Sidbury (Shropshire), Sidcup (London), Siddington (Cheshire, Gloucestershire, and Sidestrand (Norfolk). It thus follows that many settlement names are as old as river names and can also be interpreted in terms of Archaic Gaelic. The occurrence of similar names in Scotland show that this language was once common to Britain as a whole.

The interpretations proposed are in many cases only the most probable among several options. The reason for this is the notable overlap between ‘river’ words and ‘fire’ words in Gaelic. This may reflect the fact that early settlement occupied river valleys or it may reflect some deeper association of ideas.

Gaelic has àin ‘heat, light’ and àine ‘fire’ as well as ain ‘water’, an ‘water’, èan ‘water’ and aon ‘country’ (aon ‘one’ is a single defined unit). It also has éibh ‘fire’ and ab or àbh ‘river’. The compound word àbhainn ‘river’ (Avon) appears to mean ‘fire-river’. A similar overlap occurs between G. bior ‘water’ and AG. bior ‘fire’ (found in bior-fuinn ‘landmark, beacon’). The generic river name Amber might be thought to make better sense as ‘fire river’ or ‘settlement river’ than as ‘river-river’. It is possible that the fires in so many river names were not the domestic fires of hunting camps but beacons lit to organise hunts or to define territories. This ambiguity is brought out by G. dobhair which can mean ‘territory’, ‘boundary’ or ‘water’ and which is found in river names and settlement names in both England and Scotland. The first element is cognate with AG. dubh ‘deer’ and the most appropriate sense is ‘tribal hunting lands defined by water’.

The general rule is that most place-names are as old as the named elements. The assumption is that, once named, a feature or locality very seldom changes its name. The whole point of naming a place is to create a permanent identify for it. If the names of most features are as old as the features themselves, most river names must be as old as the Mesolithic. Naming a new country has always been a priority with new settlers, since names serve as a way of finding one’s way around. Elements added since the Mesolithic will obviously have later name but the landscape seems to have been densely named at an early period.

In hilly country rivers are central to settlement and in flat country they are still often used as boundaries and in both cases would be among the earliest features to be named. However, given the power of clerks to influence place-names, it is quite probable that many smaller rivers were renamed at a later date in Anglo-Saxon or even modern English – Winterbourne is one such name. A few English rivers are known to have changed their names or go today by two different names. The Asker in Dorset was once the Loders, the Swift in Leicestershire was probably once the Lutter, and the river Limen in Kent is now known as the East Rother. The Thames is also the Isis and the Granta is also the Cam. Since the new names are as obscure as the old names in most of these cases, it is possible that these rivers once had different names at different places.

The custom of giving a river the same name throughout its length may be modern. River names are among the oldest and simplest names in Britain. The first settlements in Britain were on the banks of rivers, which were easy of access by land and by water and provided drinking water and, in times of emergency, fish to eat. The word ‘river’ is also cognate with E. tribe, another pointer to the link between tribal settlement and river catchment areas.