Attested:  Ptolemy 2,3,20 Ρατε (or Ραγε);  AI iter 6 Ratas and iter 8 Ratis
milestone ... A RATIS M II  ‘2 miles from Ratae’;  RATI on a diploma (photo here)

WhereLeicester, a Roman city but probably never a fort, on the Fosse Way and the river Soar, around SK582044.  Securely located by that milestone and the AI mileages.

Name origin:  Most books accept the view of R&S that the Latin nominative form was Ratae (feminine plural), latinised from “British *rātis”, related to words in Celtic languages meaning ‘fortification’.  But there are problems ....
    Old Irish ráth or ráith can mean ‘surety, guarantee’ or ‘earthen rampart surrounding a chief's residence, fort’.  Fitzpatrick (2009) explained how the essence of a ráth (as distinct from a lis, dun, cashel, cathair, crannog, etc) was earthiness (rather than defensive strength) so that ringfort was a less-than-ideal translation.  In any case, something like 45 thousand instances are known throughout Ireland, and ráth has contributed to hundreds of Rath- place names in Ireland, plus analogues in Scotland and other Celtic areas.
    We have seen no evidence that Ratae, or any other Roman-era places with Rat- names, possessed a ráth-type structure, hardly surprisingly if that was inherently made of earth.  Also, there is no evidence that any Irish Rath- place received its name before Christian times, but that is no reason to doubt that ráth structures sprang from prehistoric traditions.
    Edmund Spenser (1596, vol 6, p 628) wrote: “There is great use among the Irish to make great assemblies together upon a Rath or hill, there to parley”.  This puts ráth firmly among the structures discussed at great length in two books by Hadrian Allcroft (1927,1930).  He showed that from deep in prehistory circular structures, which could be either or both of rings (such as henges) and cones (such as moot mounds), were associated with burials, veneration of ancestors, and tribal assemblies.  He focussed primarily on one set of words (cruc, circus, curia, church, kirk, etc) which show a sense development from places of deliberation to places of worship.
    Maybe ráth is fundamentally the same word as German Rath, the early spelling for Rat ‘council’, seen in modern words such as Rathaus, and cognate with OE rćd ‘counsel, advice’, seen in names such as Ethelred and Alfred.  Maybe all the cows and cattle-raiding in Ireland favoured sense development towards earthen cattle corrals, whereas all the trees in Germany led towards town halls.
    We have seen no convincing etymology for rath and its Celtic cognates, even by Delamarre (2003), who explains other ancient words ratis ‘fern’ and rato- ‘fortune’.  The root of Germanic Rath etc may be *redh-, derived from PIE *əre- ‘to reason, to count’ (Pokorny’s 55).  This whole area could stand more work by a heavy-duty linguist, since there are other difficulties, such as with rat (the animal).  Anyway, the best explanation for Ratae and other Roman-era Rat- names probably lies in another direction altogether.

Possibly parallel names elsewhere:
Ratiaria (a Roman naval base on the Danube, near Vidim, Bulgaria)
Argentoratum (Strasbourg Roman fort, on the Rhine)
Ratiatum (Rezé, on the Loire, see here p113 about its status as a ratis-port and vicus)
Ρατοσταθυβιου (Ptolemy, at or near Magor Pill on the Severn)
DEA RATIS is on an inscription found at Chesters on Hadrian’s Wall, with a port on the river North Tyne at NY916702 (Selkirk 1995:247,258)
Ratis (RC, Ile de Ré, off La Rochelle)
Corterate (Coutras, at a river confluence and ford, upstream of Bordeaux)
Carbantorate (Carpentras, on river Auzon, tributary of the Rhone)
Ratecorion (RC, most likely the Trent Vale fort at Stoke-on-Trent)
Argentrato (800s, Argentré on river Jouanne, leading to the Loire)
Ratisbona (869 Radasbona, now Regensburg on the Danube)
Barderate (Pliny; somewhere in the Po Valley, Liguria; possibly on the river Tarano)
Baiorate (on a Merovingian coin, struck ?580 at Béré, possibly a variant of Bairacus)
RATIN BRIVATIOM is on an inscription found at Naintré, by the river Vienne.  Briva probably  meant ‘bank’ (related to brim etc) not ‘bridge’.
Ρατομαγος (Ptolemy etc, Rouen, port on the Seine, but ROTO on Merovingian coins, etc)
Ra Mastrabalae (Avienus 300s; later Beata Maria de Ratis, at Saintes-Maries-de-la-Mer, Rhone delta).
Suarattaratae = Saurashtra in India. Pliny 6,75 also mentions Oratae

Latin ratis ‘log-raft, float, boat’ obviously jumps out as the most likely explanation for the river-port status of all these 16 places.  (We have listed them so that those whose ancient port status is best attested are at the top and those whose late date or debatable name or place are at the bottom.)  Most, though certainly not all, human settlements are near water.  However, it is striking that these Rat- places seem to be significant precisely because they were adjacent to small-boat-navigable water, rather than just being optimally sited for military control of an area and happening to have a wharf.

The etymology of ratis is “uncertain ...has been connected with Latin retae ‘trees in the bed of a stream’ and/or with ... OE rod, but neither is semantically compelling ... ‘the rowing’”, according to de Vaan (2008).  Besides being used to make boats, lots of logs go into palisade-type fortifications or murus gallicus fortress walls.  Rafts, as distinct from properly shaped boats, would have been used to send timber down-river, and radeau ‘raft’ could even refer to an island of slightly high ground in a marsh (Benoit, 1940).

In summary, it seems most likely that Ratae was named from its situation on the river Soar, at one of the reasonable limits of Roman-era water transport leading via the rivers Trent and Humber to the sea.  Irish ráth and its Celtic cognates remain interesting in their own right, but should cease being quoted as a strong argument whether Roman-era names were created in a Celtic or Germanic language.

Standard terms of use:  You may copy this text freely, so long as you acknowledge its source, recognise that it is subject to human error, and try to offer suggestions for improvement.
Last Edited: 2 October 2016