Derek Watkins has created a striking map that he has posted here:

This map taps into the place names contained in the USGS National Hydrography Dataset to show how the generic names of streams vary across the lower 48. Creeks and rivers are symbolized in gray due to their ubiquity (although the etymology behind the American use of creek is interesting), while bright colors symbolize other popular toponyms.

Lite-Brite aesthetic notwithstanding, I like this map because it illustrates the range of cultural and environmental factors that affect how we label and interact with the world. Lime green bayous follow historical French settlement patterns along the Gulf Coast and up Louisiana streams. The distribution of the Dutch-derived term kill (dark blue) in New York echoes the colonial settlement of “New Netherland” (as well as furnishing half of a specific toponym to the Catskill Mountains). Similarly, the spanish-derived terms rio, arroyo, and cañada (orange hues) trace the early advances of conquistadors into present-day northern New Mexico, an area that still retains some unique cultural traits. Washes in the southwest reflect the intermittent rainfall of the region, while streams named swamps (desaturated green) along the Atlantic seaboard highlight where the coastal plain meets the Appalachian Piedmont at the fall line.

There’s more discussion at the link, and James Cheshire, impressed by Watkins’s work, has produced a similar map (posted here) that does the same for the major rivers and streams in Great Britain. It’s a wonderful use of cartography to illustrate language use. [Dec. 2023: Cheshire has modernized his blog, and the river-names post is here, but I have provided an archived link above because the modern version has ditched the comments. Tsk.]


  1. Athel Cornish-Bowden says

    He wonders why brook is so much more common in Britain than in the USA. It is of course very much of a living word in Britain today, not necessarily associated with a name — maybe not as common as stream, perhaps, but getting there. Any river small enough to be jumped across is likely to be called a brook.

  2. It would be interesting to do the creek/brook analysis for Canada and Australia too. My unscientific observation is that “creek” is the more common in these countries as well. You might expect Australia to be closer to English usage, but I don’t think it is. On the other hand, all three countries were settled at roughly the same time, whereas I suppose most features in England acquired their names much earlier.
    Ireland might be an interesting comparison, since there are parts of Ireland where English placenames were bestowed by the Ordnance Survey in the early 19th century. There are some rivers with the first component in the name being “Owen” corresponding to Irish “abhainn”=river. However, I don’t recall either “creek” or “brook” being much used in Ireland. Perhaps someone who lives there could comment.

  3. The OED gives an explanation for the different uses of creek. The English used it to mean a narrow inlet by the sea or in a river estuary. This makes sense to me, there are creeks along the banks of the Thames in London. When explorers went to the colonies what they thought were creeks in the English sense turned out to be longer – i.e. streams or smaller rivers in their own right – and so the colonial usage added a meaning.
    creek, n.1
    Forms: α. 4 krike, 4–5 cryke, (kryk), 6–7 crike; β. 4–6 creke, (6 creake, crieque), 6–7 creeke, 7 creak, creick, 6– creek; γ. 6– crick.
    [Three types of this are found, viz. (1) crike, cryke (ī), usual in ME., (2) creke, rare in ME. (see sense 7), but common in the 16th c. (whence the current creek), and (3) crick, only since 16th c. The first corresponds to F. crique (14th c. in Littré); the second to earlier Du. krēke (Kilian), mod.Du. kreek creek, bay, and to med. (Anglo) L. creca (sometimes crecca) creek. The form crick resembles Sw. dial krik bend, nook, corner, creek, cove (Rietz), and Icel. kriki crack, nook (handarkriki armpit), but is prob. an Eng. shortening of crique, crike. In many parts of U.S. crick is the common pronunciation of creek in the sense ‘stream’. The earlier history is not known, but the word (in French also) is generally supposed to be Germanic. In sense 4 the word appears to be related to crack; in 6 and 7 there appears to be association with crook.
       A corresponding double form is seen in pike, peak, F. pic. It has been conjectured that the word is preserved in the OE. proper names Creacanford, Crecᴁanford, Creᴁanford, Crayford (in Kent), and Crecca-ᴁelád, Cricᴁelad, Flor. Criccelade, Cricklade (in Wilts); the former is impossible; in the latter crecca could not be the origin of either crike or creke, though some connexion is possible, if there were any reason to suppose that the meaning suits.]
    I. 1.I.1 a.I.1.a A narrow recess or inlet in the coast-line of the sea, or the tidal estuary of a river; an armlet of the sea which runs inland in a comparatively narrow channel and offers facilities for harbouring and unloading smaller ships.
       (The first quot. may be of more general meaning.)
    α    c 1250 Gen. & Ex. 2947 In euerilc welle, in euerilc crike [printed trike] Men funden blod al witterlike.    c 1300 Havelok 708 Hise ship‥He dede it tere, an ful wel pike, Þat it ne doutede sond ne krike.    c 1386 Chaucer Prol. 409 He knew‥euery cryke in Britaigne and in Spayne.    c 1440 Promp. Parv. 103 Cryke of watyr, scatera.    1542–3 Act 34–5 Hen. VIII, c. 9 §1 Dwellers next vnto the streme of Seuerne, and vnto the crikes and pilles of ye same from Kingrod vpward toward the City and Towne of Gloucester.    1626 Capt. Smith Accid. Yng. Seamen 17 A channell, a bay, a rode‥a crike, a riuer.
    β    1512 Act 4 Hen. VIII, c. 1 §1 The Frenchemen‥knowe‥every haven and Creke within the sayde Countie.    1571 Hanmer Chron. Irel. (1633) 155 The ship was by foule weather driven into a creick.    1622 Callis Stat. Sewers (1647) 38 Creek of the sea is an Inlet of sea cornered into the main Land, shooting with a narrow passage into some Angle of the Land, and therein stretching it self more then ordinary into the Land.    1694 Smith & Walford Acc. Sev. Late Voy. i. (1711) 39 A Creek two miles long, which is dry at Low Water, and not more than thirty foot broad.    1839 Penny Cycl. XIII. 187/2 (Faversham) The creek or arm of the Swale on which the town stands is navigable for vessels of 150 tons.    1846 McCulloch Acc. Brit. Empire (1854) I. 57 A long narrow saltwater creek, communicating with the sea at Portland Road.
    γ    1582 N. Lichefield tr. Castanheda’s Conq. E. Ind. 64 a, Foysts placed in euery Baye or Kricke to set upon him.
    b.I.1.b A small port or harbour; an inlet within the limits of a haven or port. c.I.1.c In the Customs administration of Great Britain, an inlet, etc., not of sufficient importance to be a separate Customs station, but included within the jurisdiction of another port station.
    α    1478 Botoner Itin. (Nasmith 1778) 125 Pertinentes ad havyn de Falmouth sunt 147 portus et crykes.
    β    1486 C’tess of Oxford in Four C. Eng. Lett. 7 That such wetche‥be used and hadde in the poorts, and creks.    1588 Act 1 Eliz. c. 11 §1 Conveying‥their Wares‥out of Creekes and Places where no Customer ys resident.    1642 Milton Apol. Smect. (1851) 298 He must cut out large docks and creeks into his text to unlode the foolish frigate of his unseasonable autorities.    1789 Ann. Reg. 133 A Creek in the language of the Customs, is a place included within the limits either of a head or member-port; as set out by the commissions of the Court of Exchequer; and at which officers competent to transact the coast business are stationed by order of the Board of Customs.    1863 P. Barry Dockyard Econ. 211 Between the fourth and fifth slip there is a dock inlet or creek, which might at any time be enlarged into a dry dock or basin for ships of the largest class.    1876 Act 39 & 40 Vict. c. 36 §11 Customs Laws Consolidation. The pre-existing limits of any port, sub-port, haven, creek, or legal quay.
    γ    1628 Digby Voy. Medit. 47 Besides the port is but a little cricke.
    d.I.1.d Applied to any similar opening on the shore of a lake.
       1810 Scott Lady of L. i. xiv, Loch-Katrine‥In all her length extended lay, With promontory, creek, and bay.
    2.I.2 As part of a river or river-system. a.I.2.a An inlet or short arm of a river, such as runs up into the widened mouth of a ditch or small stream, or fills any short ravine or cutting that joins the river. (This is merely an occasional extension of sense 1.)
       1577 [see 8].    1653 Walton Angler 147 A He and a She Pike will usually go together out of a River into some ditch or creek.    1671 Milton P.R. ii. 25 On the bank of Jordan, by a creek, Where winds with reeds and osiers whispering play.    1774 Goldsm. Nat. Hist. (1776) IV. 151 The otter has two different methods of fishing; the one‥by pursuing [its prey] into some little creek, and seizing it there.    1814 D. H. O’Brien Narr. Escape 109 On the banks of the Rhine‥I‥perceived a small Punt hauled into a creek.    1882 M. E. Braddon Mt. Royal I. vi. 151 He knew every tributary, creek, and eyot.
    b.I.2.b In U.S., Canada, Australia, New Zealand, etc.: A branch of a main river, a tributary river; a rivulet, brook, small stream, or run.
       Probably the name was originally given by the explorers of a river to the various inlets or arms observed to run out of it, and of which only the mouths were seen in passing; when at a later period these ‘creeks’ were explored, they were often found to be tributaries of great length; but they retained the designation originally given, and ‘creek’ thus received an application entirely unknown in Great Britain.

       1622 in H. R. Shurtleff Log Cabin Myth (1939) 155 Creeks and Swamps as they call them‥offer all aduantages to their‥enimys.    1638 in Massachusetts Hist. Soc. Coll. 3rd Ser. VI. 42 They‥are dispersed securely in their plantations sixty miles along the coast, and within the land also, along some small creeks and rivers.    1674 Pennsylv. Archives I. 34 On the East-side of a Small Creeke or gutt on this side the Single-tree.    1748 F. Smith Voy. Disc. N.W. Pass. I. 132 Called Ten Shilling Creek, but not properly, it being a Branch of the great River.    1748 Washington Jrnl. 25 Mar., Left Cresaps and went up to ye mouth of Patersons Creek [a tributary of the Potomac].    1793 J. Hunter Hist. Jrnl. Port Jackson xxi. 516 In the afternoon a creek obliged them to leave the banks of the river, and go round its head, as it was too deep to cross.    1820 S. Marsden Let. 8 May (1932) iv. 249 There are two warm springs opposite each other, one on each bank of the creek, about ten feet above the level of the fresh water which runs between them.    1836 Backwoods of Canada 64 Besides numerous small streams, here called creeks, two considerable rivers‥find an outlet.    1840 N.Z. Jrnl. I. 292/1 A fine creek of water‥runs down from the mountains‥and wastes itself in the salt water.    1852 G. B. Earp Gold Colonies Australia viii. 131 A ‘creek’, in New South Wales jargon, means a water-hole in the interior, and not an arm of the sea, as we understand it.    1857 R. Paul Lett. Canterbury iv. 65 In the Australian Colonies, as in America, brooks are called creeks.    1879 D. M. Wallace Australas. ii. 25 The drainage of the interior is effected by numerous creeks and water⁓courses which only run after periods of rain.    c 1848 in H. Watterson Oddities S. Life & Char. (1883) 69 ‘You see that krick swamp?’ asked Suggs.    1963 B. Pearson Coal Flat iv. 59 In a little clearing in the‥young bush by the bank of the creek was Mrs Seldom’s little house.
    c.I.2.c Phr. up the creek: (a) in a tight corner, in trouble; spec. pregnant; (b) crazy, eccentric. slang.
       1941 A. Miller in Kozlenko 100 Radio Plays 22/2, I mean that if I’m killed you’re up the creek.    1943 Hunt & Pringle Service Slang 68 Up the creek, lost, either on patrol or during a night out; off one’s course.    1945 Penguin New Writing XXIV. 32 Lord, it’s my mortar lance corporal. If he breaks we’ll be right up the creek.    1960 H. Pinter Caretaker iii. 71 You don’t know what you’re doing.‥ You’re up the creek! You’re half off!    1961 H. E. Bates Day of Tortoise 52 I’m in trouble. I’m going to have a baby.‥ I’ve had it. Good and proper. I’m up the creek.    1961 J. Heller Catch-22 (1962) viii. 78 You really are up the creek, Popinjay.    1963 E. Lambert Drip Dry Man x. 51, I know a girl who thinks her bloke may have put her up the creek.    1969 I. Kemp Brit. G.I. in Vietnam vi. 134 ‘You okay?’ asked Donovan.‥ ‘I thought you were properly up the creek.’
    3.I.3 transf. senses akin to 1. †a.I.3.a Applied more widely and loosely to any narrow arm or corner of the sea. Obs.
       1635 N. Carpenter Geog. Del. ii. vi. 87 The Adriatic Sea in the inmost creeke neere Venice swels neere foure foote in hight.    1652 Needham Selden’s Mare Cl. 333 Jersey, and Garnesey‥situated within that Creek of Sea which is made by the shore of Bretaign on the one side, and that of Normandie on the other.
    b.I.3.b A narrow corner of land running out from the main area; a narrow plain or recess running in between mountains. Cf. cove.
       1649 W. Blithe Eng. Improv. Impr. (1653) 56 Certain Creeks or corners of Land running into the up-lands.    1669 Worlidge Syst. Agric. xi. §3 (1681) 233 To How the several Creeks, Corners, and Patches of your Land.    1856 Stanley Sinai & Pal. ii. (1858) 136 The plains which run into the mountains are the creeks into which they [the Bedouins] naturally penetrate.
    II.II †4.II.4 A cleft in the face of a rock, etc.; a crack, fissure, chink, crevice, cranny. Obs.
    α    1375 Barbour Bruce x. 602 Thai clam into the crykis swa, Quhill half the craggis thai clummyn had.    c 1375 Sc. Leg. Saints, Blasius 43 A kryk in to a crage he hade, & þare his dwellinge has he mad.
    β, γ    1570 Levins Manip. 54 A creke, crick, fissura.    Ibid. 120 A crick, rima.    1635 R. Brathwait Arcad. Pr. 179 To wals and portels would he lay his eare, Through creeks and crannies too, that he might hear‥desir’d applause.
    5. a.II.5.a A narrow or winding passage penetrating the interior of any place and passing out of sight; an out-of-the-way corner. to seek creeks: to seek a hiding-place. Obs. or dial.
       1573 Tusser Husb. (1878) 108 Tom Piper hath houen and puffed vp cheekes, If cheese be so houen, make Cisse to seeke creekes.    1582 T. Watson Centurie of Loue xcv. (Arb.) 131 A Labyrinth is a place made full of turnings and creekes.    1590 Shakes. Com. Err. iv. ii. 38 One that countermands The passages of allies, creekes and narrow lands.    1629 Chapman Juvenal v. 15 Is no creek void?    1681 Cotton Wond. Peak 52 The Cave‥stretching itself‥As if (past these blind Creeks) we now were come into the‥Mountains Womb.    1750 Gray Poems, Long Story 62 Each hole and cupboard they explore, Each creek and cranny of his chamber.    1808–25 Jamieson s.v. Crykes, ‘Creeks and corners’ is still a common phrase.    1878 Mrs. H. Wood Pomeroy Ab. (ed. 3) 112 We‥looked in every crick and corner for it.    1883 G. Allen in Colin Clout’s Calendar 65 To fill up all the cricks and corners between other plants.
    b.II.5.b fig. A nook, a hidden or secret corner.
       1577 tr. Bullinger’s Decades (1592) 341 And so must sounde doctrine keepe all the faithfull‥in their duetie and quiet concorde, without creake or creauise.    1587 Fleming Contn. Holinshed III. 1296/1 Throughlie view the hidden and couered creeks of our minds!    1614 J. Day Day’s Festivals (1615) 261 There is not a creeke or crany in the World, but seemes to bee fraught with it.    c 1620 Z. Boyd Zion’s Flowers (1855) 91 The crooked creekes Within my heart.    1715 M. Davies Athen. Brit. i. 249 Jesuitical Creeks and Corners of Superstitious Romanism.
    †c.II.5.c Applied to the two cavities of the heart.
       1621–51 Burton Anat. Mel. i. i. i. iii, The Ventricles, Caules, Kells, Tunicles, Creeks, and parts of it.    Ibid. i. i. ii. iv, This heart, though it be one sole member, yet it may be divided into two creeks right and left.
    †6.II.6 A turn, a winding, as of a river or crooked way. Also fig. Obs.
       1592 Davies Immort. Soul xv. 4 As Streams, which with their winding Banks do play, Stopp’d by their Creeks, run softly thro’ the Plain.    1596 ― Orchestra, I love Meander’s path‥Such winding slights, such turns and cricks he hath, Such creaks, such wrenches, and such dalliance.    c 1653 England’s Alarm in R. Bell Collect. Anc. Songs 1857 Painted harlots which they often meet At every creek and corner of the street.    1671 J. Flavel Fount. of Life vii. 20 In every Creek and turning of your lives.    a 1680 Charnock Attrib. God (1834) I. 648 He hath a prospect of every little creek in any design.
    †7.II.7 fig. A crooked device; a trick, artifice, contrivance. Obs.
       The early date of this sense makes its history and position doubtful.
       c 1386 Chaucer Reeve’s T. 131 (Ellesmere MS.) They wene þat no man may hem bigile‥The moore queynte crekes [so 4 MSS.; Harl. knakkes] that they make The moore wol I stele.    a 1626 W. Sclater in Spurgeon Treas. Dav. Ps. cxvi. 6 Without those slights, or creeks of carnal policy, for which men are in the world esteemed wise.
    III. 8.III.8 attrib. and Comb., as creek-bed, creek-hole, creek-side, creek-timber (U.S.). creek-bottom U.S., level ground beside a stream or brook; creek-gum, the Australian river red gum, Eucalyptus camaldulensis.
       1852 Southern Lit. Messenger XVIII. 314/1 Ahead of us‥is a creek-bed, fringed far back into the hills with tree and shrub.    1859 K. Cornwallis New World I. 111 A creek-bed ran parallel with the road.    1936 I. L. Idriess Cattle King iv. 30 Often you can dig in a dry creek-bed and obtain soakage water.    1965 M. Shadbolt Among Cinders xx. 194 We walked‥following a creek-bed.
       1822 J. Woods 2 Yrs. Eng. Prairie 224 On the creek bottoms, coffee-berry, poplar, pecon, white walnut.    1857 F. L. Olmsted Journ. Texas 81 The soil of the creek-bottoms bears good cotton.
       1898 Morris Austral Eng. 178 Creek Gum.    1930 V. Palmer Men are Human x. 89 The sun, visible through a gap in the creek-gums, hung fixed in the sky.
       1748 F. Smith Voy. Disc. N.W. Pass. I. 145 The Pieces of Swamp between this Channel and the Creek-head.
       1577 B. Googe Heresbach’s Husb. iv. (1586) 173 In the bankes and sides of these Ponds, you must have Bushes and Creeke holes, for the Fish to hide them in from the heate of the Sunne.
       1879 Whittier St. John xvii, From island and creek-side Her fishers shall throng.
       1790 J. Backus Diary in W. W. Backus Mem. Backus Fam. (1889) 93 Explored the bottom to the Creek timber land.    1836 D. B. Edward Hist. Texas 36 We find‥[the] prairie‥relieved by creek timbers and solitary groves.
    Hence ˈcreekward a., towards a creek.
       1887 C. C. Abbott Waste-Land Wand. iii. 85 Kept a creekward course until out of sight.

  4. That’s caused a bit of a bottleneck at the start of the comments, so feel free to reduce the OED quote to the bolded bit, Language. I left the whole thing only because I enjoy reading the cited examples. Unfortunately you can’t link to Oxford publications (bastards).

  5. Interesting how both Northern and Southern Appalachia are picked out and distinguished from each other.

  6. I can’t see that the British river-names map says anything useful, certainly for England (although the Scottish divisions look interesting): while in the US there is clearly a variety of synonyms used for similar sorts of running water, in England it’s pretty much river = large to medium, brook = medium/medium-small and stream = small, and only “beck” (= German Bach) and “gill” as regional variants (for a brook and a stream, respectively, mostly in the old Danelaw).
    What would have been more interesting is a map showing the linguistic roots of the river names themselves, which would reveal just how many English rivers, even in the east, have Celtic names (Avon, of course, and Ouse, and Exe and Ash [both “isca”, “water” as in “uisge”/whisky] among many others).

  7. I don’t recall either “creek” or “brook” being much used in Ireland. Perhaps someone who lives there could comment.
    Me neither. Checking the Irish Placenames Database for the four corners of the island and, apart from a few “X Water”, it’s all “X River” or “River X”.

  8. “X Water” surely marks places settled by Scots.

  9. We’ve got plenty of “brooks” here in E Anglia; I don’t remember living near one in Scotland. Watercourses smaller than rivers and bigger than burns were often “water”, as in “Water of Milk” or “Kirtle Water”.
    Mind you, why no “bourne” in the south of England?

  10. Isn’t the Bourne River (or is it River Bourne?) in the south of England?

  11. I think the Bourne flows into the Serpentine, the lake in Hyde Park. Like most of the old rivers of London it’s enclosed in a tunnel. When I was young, it used to leak into the basement of Elliot’s shoe shop on the corner of Westbourne Grove.
    There are lots and lots of -bourne place names in southern England. Not just Eastbourne and Bournemouth, but Pangbourne is on the Thames, near Reading. There’s Melbourn and Bassingbourn in Cambridgeshire, near where my grandmother grew up (Kneesworth) – they can’t be so far from dearie’s house.

  12. My remark was unclear – I meant why didn’t the mapmaker include an entry for “bourn”/”bourne”, the word being so common. It seems a strange omission.
    P.S. Is “bourn”/”bourne” common anywhere in the US? (Or Canada, ….)

  13. There’s also Bournville, a town made entirely out of chocolate, but I like to think of the Midlands as ‘up north’.

  14. Although the Cadburys were Quakers (like the Frys, Roundtrees and all the other chocolate makers in Britain), Bournville has its own Serbian Orthodox church. The interior surface of the dome is entirely covered with milk-chocolate buttons.

  15. We have forks and washes galore in Utah, but also a number of names not shown on Derek’s map – gulches, canyons, hollows, and even a small group of “waters” (all located in Diamond Fork drainage and called 1st, 2nd, 3rd, 4th, 5th and 6th Waters resp.)
    Forks here were usually named after early settlers who’ve been granted rights to built logging roads or livestock trails there, so it’s most commonly “Last Name[‘s] Fork”. But sometimes their names get nested (West Fork Blacks Fork, yes, that’s a name of four parts in a region where many names are similarly long-winded), or sometimes “nominees” didn’t want their names used (like when Apostle Smoot built a logging road to supply timbers for the Salt Lake Tabernacle, he insisted that the creek be called Church Fork, for he’s done the work on the Mormon Church’s behalf)
    How do you classify “Left Fork Bear River” btw – as a river or as a fork 😉 ?

  16. “Pangborn” has always suggested the pain of childbirth to me, as if the original person of that name was someone whose mother had a particularly difficult time of it.

  17. It’s from a Berkshire place name recorded in 844 as Pegingaburnan, meaning ‘stream of the family or followers of a man called Pæga’ according to the Oxford Dictionary of English Place-Names, which adds: “The river-name Pang is a ‘back-formation’ from the place-name.”

  18. I note here that I am a great fan of the works of Edgar Pangborn, especially the novel Davy and the short-story collection Still I Persist In Wondering, which in full continues “whether folly must always be our destiny.”
    One hopes not.

  19. David Marjanović says

    Watercourses smaller than rivers and bigger than burns were often “water”, as in “Water of Milk” or “Kirtle Water”.

    Tiroler Ache and lots of others in -ach, all of them grammatically feminine. There it is, the other Indo-European “water” word, preserved for long enough to pass through the High German consonant shift.

  20. Oh yes, and that’s not the only case: Aachen is Aquisgranum, of course. Then there’s Die Aumühle, a village on the Helme where my mother was brought up in the 1920s, consisting at the time solely of the mill itself. (She was born in Heringen, a few kilometers to the west). I had a claim to ownership of this mill at one time as the sole survivor of my mother’s maternal line, but no interest in pursuing it. I see there is more of a village there now; there was a railway station there (probably mostly for freight) which is why the place got onto maps at all.

  21. @David Marjanović:

    There it is, the other Indo-European “water” word

    I wouldn’t call *akʷ-ā- the ‘other’ PIE word; after all, it’s the base word in the Romance branch and survives in Northern Germanic as well (Swedish ‘å’, for example).
    Closer to ‘Ache’ you still have Dutch ‘aa’ and Frisian ‘ie’, not counting dialectal forms.

  22. David Marjanović says

    Oh, yeah. Au(e) means “forested floodplain” and occurs in lots and lots of placenames.
    And I notice only now that this is actually the native Germanic version: apparently, the /kʷ/ has undergone Verner’s law, turning into /ɣʷ/ and then /w/, which then formed a diphthong with the preceding /a/ and therefore escaped the North Germanic and Continental West Germanic shift to [v].
    So, Ache must be Romance just as much as Aachen and has only undergone the High German consonant shift. That explains the geographic distribution of this element – it only occurs south of the Danube and west of the Rhine, AFAIK.

  23. An Au is certainly a floodplain, but I don’t think it has to be forested: it can be a water meadow just as well.
    Back to Tolkien’s “English and Welsh” for a moment: it was his theory that [w], [ð], and [θ] were preserved in those two languages but not in Frisian, Dutch, German, Breton, or Gaelic because of a Sprachbund effect. (The preservation of the last two in Icelandic and the last in Faroese is straightforwardly accounted for by their extreme isolation.)
    Likewise, both Old English and Old Welsh had a doubled present tense: the inherited am, is, are forms and a second set of forms based on be, the latter having (uniquely among the related languages) both future and consuetudinal meaning. This apparently spread from Welsh to English by semantic extension.

  24. David Marjanović says

    An Au is certainly a floodplain, but I don’t think it has to be forested: it can be a water meadow just as well.

    That meaning is probably extinct. It occurs in the psalms: Er weidet mich auf grüner Aue

    it was his theory that [w], [ð], and [θ] were preserved in those two languages but not in Frisian, Dutch, German, Breton, or Gaelic because of a Sprachbund effect

    Yes… but I wonder if this is simply to blame on shared splendid isolation. First of all, West Flemish retains [w]. Secondly, I think neither the Other Continental West Germanic change from [w] to [v] or [ʋ] nor the Gaelic acquisition of [v] as an allophone of /w/ should be expected to have reached south of Hadrian’s Wall… Thirdly, neither the Continental West Germanic change of [θ] and [ð] to [d] nor the Gaelic disappearance of [ð] and change of [θ] to [h] should be expected to have reached south of Hadrian’s Wall. Anything that happened in Breton can, I’m sure, be blamed on French.

    Likewise, both Old English and Old Welsh had a doubled present tense: the inherited am, is, are forms and a second set of forms based on be, the latter having (uniquely among the related languages) both future and consuetudinal meaning. This apparently spread from Welsh to English by semantic extension.

    That can only apply to the meanings. The existence of two verbs must be older. German has bin and bist, and the Slavic languages have the same pattern as modern English.

  25. It’s the existence of two competing tenses that are both formally Presents, plus the specific meanings, that is particular to English and Welsh. Be in Indo-European languages is the result of fusing up to three different roots. The *esmi- ‘be’ root gives E am, is, are, G ist, L esse, es, est, and is the dominant verb in Balto-Slavic and Indo-Iranian. With loss of the first vowel, it gives us L sum, sumus, sunt and G sind; this form was lost early in E.
    In Germanic, however, this verb early merged with *wes- ‘remain, dwell’, which gives E was, were, G war, warst, waren, wart. It retains its older meaning in Skt vasati ‘he stays’ and L Vesta, possibly also Gk Hestia, the hearth goddess, the indweller.
    Finally, and independently in different families, *esmi-, possibly combined with *wes-, merged with the descendant of bh(e)u- ‘become’, to give E be, been, being, G bin, bist, L fui, fuisti, fuit etc. As EtymOnline says, be is “a collection of semantically related paradigm fragments […] an accidental conglomeration [of forms] from the different Old English dialects”; the same is true mutatis mutandis of its equivalent in the other Western IE languages.

  26. David Marjanović says

    It’s the existence of two competing tenses that are both formally Presents, plus the specific meanings, that is particular to English and Welsh.


    With loss of the first vowel

    (Incidentally, that’s probably best explained as e-grade vs. zero-grade: *h1es- vs. *h1s-, *h1s-o-m vs. h1es-ti. A few laryngeals can make a lot of things fall into place.)

  27. Sure. But “zero-grade” = “loss of a vowel”, no?

  28. David Marjanović says

    No… there’s no way to say any of the ablaut grades is basic and the others derived, as far as I know.

  29. John Cowan says

    […] English rivers, even in the east, have Celtic names (Avon, of course, and Ouse, and Exe and Ash [both “isca”, “water” as in “uisge”/whisky] among many others).

    I’m not so sure about Ash (there are two in England); I don’t see any evidence that that isn’t plain English. But the Exe, the Axe, and the two Esks in England as well as the Usk in Wales are definitely all < isca; I’m not so sure about the many Esks in Scotland.

    Uisge, as well as Ouse, however, really are from PIE *wed-. However, isca is not ‘water’ but something like ‘abounding in fish’, according to various sources. In particular, the Usk in Welsh is the Wysg < Proto-Celtic *eskos ‘fish’. This is cognate with Latin piscus, which itself appears in Welsh as pysg ‘fish’ (singulative pysgod). It’s typical of people grading up to an elite language to keep some very basic words as well as some syntax, and in post-Roman times British Celtic dominated British Latin, especially among the East British refugees from the English.

    I got onto this because I was looking at the etymology of Devon, which is from the Latin name for the locals, the Dumnonii. Exeter, indeed, was called Isca Dumnonii at the time, as distinct from Isca Augusta, Caerleon-on-the-Usk. Dumnonii, indeed, looks very like Cordwainer Smith’s Damnoni, tall pale hominids of Earth descent who are masters of architecture and technology. And who used to be the big cheeses in Exeter? The Guild of the Cordwainers, of course.

    But anyway, what is this Dumnonii? Apparently from Proto-Celtic *dubnos, which can mean ‘world’ or (as I choose to assume in this case) ‘deep’. So who were the Dumnonii in the view of their fellow Celts? The Deep Ones. From which we learn that they are to be found not only off Innsmouth, but ….

    (Y’know, Athel Cornish-Bowden hasn’t been seen around here for a while, and he’s a Devonian. Do you suppose his feet are getting … webbed? Iä! Iä!)

  30. Latin fish is piscis. Different stem suffix, same root.

  31. PlasticPaddy says

    I think some of your derivations are a bit fishy.
    British Isca is not related to piscis. The Insular Celtic cognate to piscis was *ēsko-, which with the regular diphthongisations of ē gave Old Irish íasc (gen. éisc) and Proto-British *oesk- > Welsh wysg. Isca goes rather with Old Irish uisce and is from *u(d)skio- > Proto-British *üskjo- > *isko- (since lost in the currently extant British languages, where the ‘other’ word for water is used: Welsh dŵr (Old Welsh dwfyr), Breton dowr, Old Irish _dobur) = Irish tobar “well”

  32. David Marjanović says

    No… there’s no way to say any of the ablaut grades is basic and the others derived, as far as I know.

    Yes… while there’s no way to call e- or o-grade more basic than the other, the zero-grade is derived from both by a synchronic rule: vowels preceding any underlyingly accented morpheme are lost in the last common ancestor of the extant IE languages, vowels preceding the actual surface stress are lost in Proto-Indo-Anatolian. It’s on somewhere, I’ll go look for it.

    Isca Dumnonii

    Surely Dumnoniorum?



  33. David Marjanović says

    Found it: this conference handout and references therein.

  34. David Eddyshaw says


    Yes, if Wikipedia is to be believed. Seems an odd name for the place, mind. I’d have thought the obvious thing about Dover (especially coming from the European side) is that it suddenly isn’t water any more. Presumably the town takes its name from the river Dour.

    My ancestors don’t seem to have been very imaginative when it came to river names. “Water.” I ask you.
    (Actually, the Kusaasi are no better. As far as I could discover, the only local name for the White Volta is “River.” Admittedly there’s not a lot of scope for confusion there.)

  35. Water?

    Who said anything about water?

    Anyone can see that the root in question refers to much stronger drink.

    So much more probable etymology would be:

    Exeter – “Whisky Fortress”

    Isca Augusta – “August whisky”

    Isca Dumnonii – “Deep whisky”


  36. PlasticPaddy says

    A small correction. Tobar seems to be related to the other words but the correct Irish descendant is dobhar, This is not common, except in place names.

  37. John Cowan says

    Piscus, piscis, Dumnonii, Dumnioniorum, let’s call the whole thing off. Or as a musician friend of mine used to sing, “Carmina, burana, carmina, burana, let’s call the whole thing Orff.” A few false quantities there, but fortunately he did not have to expiate them in tears and blood.

    I think some of your derivations are a bit fishy.

    ~~ snicker ~~

    So is the whole story. Although of course there’s more than one point of view, and for some humans “Iä! Iä! Cthulhu fhtagn!” is pretty much like “Ave Maria gratia plena Dominus tecum” for others.

    Isca goes rather with Old Irish uisce

    Well, that’s it: can we be sure which one, ‘water’ or ‘fish’, the Romans mangled into isca? I’m not sure we can: this is not a matter of regular sound-change.

    Anyone can see that the root in question refers to much stronger drink.

    Indeed. Having observed that gin and water, whiskey and water, brandy and water, rum and water, etc. all intoxicate, we may apply Mill’s methods of induction to see that the intoxicant must be the common ingredient, namely water. “In water does all life begin.” —the Orange Catholic Bible

  38. January First-of-May says


    Not to be confused with Carmarthen-on-the-Usk, the town where the School in Carmarthen is said to be located.

  39. The name of great builders of Cordwainer Smith’s future was spelled “Daimoni,” which has a transparent origin unrelated to the Celts.

  40. John Cowan says

    ~~ curses ~~

  41. John Cowan: actually, Welsh “pysg” probably derives from the Latin accusative form “piscem” rather than the nominative form “piscis”: I think it was Kenneth Jackson who first pointed this out. Oh, and “pysgod” is the plural of “pysg”, not its singulative: “pysgod” does itself, however, have a singulative form, “pysgodyn” (If any Welsh speaker among Hatters could explain downthread what the exact semantic nuance is between “pysg” and “pysgodyn”, that would be very much appreciated!).

    (Inasmuch as the Welsh -od ending derives from Latin -atu(m), I have occasionally wondered whether British Romance might not have had both */piske/ and */piskatu(m)/, whose semantic relationship to one another would have been parallel to the Spanish cognates “pez” versus “pescado”: in Spanish these are not a singular/plural noun pair, of course, but “pez” is unambiguously a singular count noun, whereas “pescado” can be a non-count noun, and thus it is easy -well, for me at any rate-to imagine that Brythonic Celtic or British Romance might have re-interpreted the distinction – between */piske/ on the one hand and */piskatu(m)/ on the other- as being a simple singular/plural distinction. Incidentally, if this re-analysis took place in British Romance, there is nothing odd in assuming the singular/plural pair to have been borrowed directly into Brythonic: there is at least one Latin loan whose singular and plural forms -the latter unambiguously its Latin form, not a Brythonic one!- were both borrowed into [at least one dialect of] Brythonic).

  42. John Cowan says

    actually, Welsh “pysg” probably derives from the Latin accusative form “piscem” rather than the nominative form “piscis”

    I more or less took that for granted, as in the Romance languages (with the exception of the surviving French nominatives, to which should be added the names Charles, Georges, James.)

    Thanks for the corrections about singular, plural, and singulative. I wonder if plain pysg might be obsolete: there is no (English) Wiktionary entry (there is one for pysgod), and poking around in various Welsh dictionaries online didn’t help either.

    Note also the semantic emptiness of -ate in English: there is no particular reason why we use the verb separate instead of *separe on the one hand, and prepare instead of *preparate on the other.

    at least one

    Tell, tell!

  43. David Eddyshaw says

    The Geiriadur Prifysgol Cymru says pysg can be singular or plural, and suggests that it is borrowed respectively from piscis and pisces.

    I don’t think it’s actually used now (ə may know better.) Even Jonah’s fish in the 1588 Bible is a pysgodyn. It seems to turn up in old hymns.

    What’s with the idea that it’s borrowed from piscem rather than piscis? and how can you tell? Welsh singular forms usually reflect Brythonic nominatives, and I can certainly think of Latin loans borrowed in the nominative, like lleidr, pl lladron.

    (And what’s the mystery Latin loan borrowed in sg and pl?)

  44. PlasticPaddy says

    Back to tobar/dobhar. Although the modern sense of dobhar is “torrent”, it comes front a PIE root beginning with d and meaning deep (same as other Celtic dover/Dour words). Tobar comes from a root word beginning with b and meaning “to boil” (like Brunnen discussed on another thread). The t in tobar is traced to an earlier d but that was an affix od-bur. So I am sorry for bringing this word in.

  45. David Marjanović says

    A three-way distinction between a mass noun, a singulative singular and a singulative plural is attested elsewhere: Lameen once had a post on Algerian Arabic bəxsis “figs”, bəxsisa “one fig”, bəxsisat “countable figs”.

    German actually has this with “fish”: Fisch, der/ein Fisch, Fische – the mass noun is marked by the absence of an article.

  46. Stu Clayton says

    @JC: Having observed that gin and water, whiskey and water, brandy and water, rum and water, etc. all intoxicate, we may apply Mill’s methods of induction to see that the intoxicant must be the common ingredient, namely water.

    That does not follow from any of Mill’s methods, in particular not from the “direct method of agreement” which you seem to be invoking:

    # If two or more instances of the phenomenon under investigation have only one circumstance in common, the circumstance in which alone all the instances agree, is the cause (or effect) of the given phenomenon. #

    The phenomenon is “intoxication following the consumption of a beverage”. The instances given by you of the phenomenon have as a circumstance in common not only “has consumed water”, but also “has consumed an intoxicant”: gin, whiskey etc.

    The “method of difference” can be applied to other instances to show that “has consumed water” is not an effect, cause, or necessary part of the cause of the phenomenon:

    # If an instance in which the phenomenon under investigation occurs, and an instance in which it does not occur, have every circumstance save one in common, that one occurring only in the former; the circumstance in which alone the two instances differ, is the effect, or cause, or a necessary part of the cause, of the phenomenon. #

    One reason why the method of agreement is problematic is that there is a universal quantifier lurking behind “two or more instances”. Nevertheless it is often applied in everyday life. And you can induct over induction, which is a good sign (one requirement for universal validity of a generalizing procedure is that it can be applied to itself). I induce from your specific application of it that you have not been consuming only water today. Perhaps coffee as well ?

  47. David Eddyshaw says

    A three-way distinction between a mass noun, a singulative singular and a singulative plural is attested elsewhere

    In fact, it’s often said that with Welsh nouns which form the sg by adding -yn or -en to the plural (of which there are a good number, especially but by no means exclusively names of trees and plants) the “plural” is “really” a sort of collective; sometimes this is plausible enough (dillad “clothes”, dilledyn “item of clothing”), but in general I think this is just a confusion of form and function, and an unwillingness to believe that any language could possibly be so unlike English as to dare to form the sg from the pl: plant (sg plentyn) is no more a “collective” than “children”; sêr (sg seren) is no more a collective than “stars”, and llygod (sg llygoden) is no a more collective than “mice.”

    Moreover, though a great many tree names are indeed like derwen “oak” pl derw, not all are: afallen “apple tree” has the bog-standard pl afallennau. If there’s any subtle semantic difference between pluralisation of oak trees and pluralisation of apple trees it’s lost on me. And the plural of our totemic animal, notorious for its lack of individual gumption, is defaid, as opposed to one dafad.

    It seems entirely plausible, granted, that historically the forms that correspond to -yn/-en singulars were collectives (plant was, for certain), but synchronically this is just another way that Welsh has elaborated its number-marking system in nouns as much as possible to confuse the English.

  48. Stu Clayton says

    The Welsh have introduced a Singulative Unintuitive Gap.

  49. David Eddyshaw says

    Come to think of it, it’s SUBTRACTIVE MORPHOLOGY! Yay! The Way of Deletion explains all!

    [Edit: Bah! Stu beat me to it while the electrons were en route from Wales.]

  50. Stu Clayton says

    It was not completely fair of me. I heard them bleating the message as they were herded through Cologne, so I passed it on by direct dialing. Routed by the route.

  51. John Cowan says

    We once had the same system in English, but fishes has been fading out and replaced by fish for a long time: the 1898 OED2’s last citation of the former is 1842, and its first citation of the latter is 1400.

    Sindarin, which is modeled on Welsh to some extent, has a systematic pattern with an i-mutated distributive plural and a suffixed collective plural: nogoth ‘Dwarf’, noegyth ‘Dwarves’, nogothrim ‘the Dwarf-folk’; orch ‘orc’, yrch ‘orcs, some orcs, des orques’, orchoth ‘the Orcs’; êl < elen ‘star’, elin ‘stars’, elenath ‘starry host’; Remmirath ‘the Pleiades’ (lit. net-jewel-PL.COLL). The suffix -rim is applied to persons, -hoth the same but derogatory, and -ath to other nouns. (Note: The circumflex marks a non-phonemic overlong vowel in monosyllables.)

    (In any case, I thought the totem animal was the draig, or even the cenhinen.)

  52. David Eddyshaw says

    I thought the totem animal was the draig

    That’s what we want you to think. (Incidentally, another loan from a Latin nominative.)

  53. David Marjanović says

    Fishes has made a roaring comeback since 1898 as “different species of fish”.

  54. On Welsh singulatives: as far as my limited knowledge of the sound-changes goes, some nouns in the Brythonic languages should really have ended up without a singular-plural distinction by sound change alone (where most of the case and number inflections had the same number of syllables and none of the final-syllable inflection, subsequently lost, triggered a vowel change in earlier syllables). Of course most of these promptly borrowed a plural suffix from other types where that was inherited in order to maintain the singular/plural distinction (while case distinctions were abandoned). But perhaps words more likely to be used in the plural solved the same problem by generating a singulative instead, borrowing a pattern established by true collectives? Sêr would seem to be one where this could have happened.

  55. David Eddyshaw says

    Good thought. It would indeed account for a lot.

    Plant actually is derived from a collective (from Latin planta.) “Mouse” should go sg llyg pl llygod if it followed the common pattern of Welsh plurals preserving the oblique stem of Brythonic imparisyllabics (stem *lukōt-) so I imagine it’s remodelled by analogy with pysgod. The “proper” sg llyg means “shrew.”

  56. fishes has been fading out and replaced by fish for a long time
    I don’t care about religion but I do care about painting and it’s quite irritating that various versions of the Miraculous Draught of Fishes are now known to some people (persons) as a “Miraculous Catch of Fish,” as if draught of fishes is too hard for anyone studying Duccio, Raphael or Rubens to understand.

    Fishes is common for groups of different types of fish (a bunch of salmon plus some herring and a few cod, for example), something like ‘moneys’. “Luca Brasi sleeps with the fishes” apparently comes from the film of The Godfather but not the book, which has the prosaic “The fish means that Luca Brasi is sleeping on the bottom of the ocean.” .

  57. John Cowan, David Eddyshaw: The mystery Latin word borrowed as singular and plural is…[INSERT DRAMATIC SOUND OF CYMBALS HERE]… CORPUS, plural CORPORA, both of which have reflexes in Welsh: “corff, corffor”: whereas both mean/meant (?) “body, corpse”, “corffor” also has the additional meaning of “bodily form”, which may be a semantic development from a plural noun: according to my sources “corff” also once had “corfforoedd” and/or “corfforion” as plurals (today the plural is “cyrff”), both of which seem to derive from the Latin/Romance plural CORPORA with the addition in both instances of a nice native Welsh suffix. Which implies that CORPORA/corff must have been understood as a plural within the history of late Brythonic/early Welsh.

    David Eddyshaw: phonologically “pysg” would have been the reflex of Latin “piscis” or “piscem”, but Kenneth Jackson (LANGUAGE AND HISTORY IN EARLY BRITAIN, 1953, footnote, page 574), in arguing that evidence indicates that short /i/ should have triggered umlaut (AKA i-mutation) in Brythonic, points out that the absence of i-mutation in some ten Latin/Romance loanwords in Brythonic, all of which ended in /is/ (with a short /i/!) in the nominative, can be explained if we assume that the Latin/Romance loanwords were borrowed in the form of the accusative (or ablative, which for most nouns seems unlikely). If this is so then the natural assumption must be that “pysg” likewise derives from PISCE(M) rather than PISCIS.

    More broadly, Harald Haarmann, in his 1970 book DER LATEINISCHE LEHNWORTSCHATZ IM KYMRISCHEN (Bonn), points out (page 156) that most borrowed nouns in Welsh derive from the accusative: tellingly, in the case of the large minority of nouns deriving from the nominative, in Old/Middle Welsh the accusative form typically coexisted with this nominative (as doublets: despite their etymology, there is no evidence that such doublets were grammatically differentiated according to case). Thus while Welsh “draig” certainly derives from the Latin nominative (DRACO), in earlier Welsh there was a doublet deriving from DRACONEM.

    Incidentally, I find it amusing that the symbol of Wales (well, to us outsiders!), Y DDRAIG GOCH, consists of a Latin/Romance noun and adjective: as for the Y, it is certainly a native Celtic form, but my personal hunch is that its use as a definite article in Brythonic languages (including Welsh) is due to the influence of British Latin/Romance.

    Oh, and finally, on “fish” and “fishes”: I sense that in many parts of the English-speaking world speakers feel more at ease in pluralizing “fish” as “fishies” than in having to choose between “fish” and “fishes” (without, be it noted, feeling the need to generalize “fishy” in the singular). Can any hatter(s) out there confirm or deny?

  58. “Fishies” is baby talk. Not that there is anything intrinsically wrong with that; Ernie and Bert, whom I respect tremendously, said “fishies.”

  59. Lars (the original one) says

    The singular / plural / collective distinction is not really that different from count nouns with their plurals vs mass nouns. Danish has a fair number of nouns that can be used as both count and mass, among them most tree names and game animals (as well as the culturally defining øl) and we even have dialects where the common/neuter distinction (in articles) has been regrammaticalized as count and mass genders.

    Actually I think that’s widespread in SAE, but I’m most sure about the facts in Danish. Der er meget eg i den her skov = ‘There’e a lot of oak in this forest’ (mass, formally singular) versus Der er mange ege i den her skov = ‘There are many oaks in this forest’ (count, plural) — and you can form the explicitly count compound noun egetræ = ‘oak tree’ to avoid ambiguity. And as can be seen, this all works in English too. (French has this thing about using du/de la for some things and de/des for others, I was unable to make sense of it in high school but maybe it’s related to this).

    So Welsh has a slightly different take on this and the grammars have a neat name for the explicit count singular because it’s not ‘just’ a compound — but give the rest of us a few hundred years and we might do the same thing.

  60. John Cowan says

    The instances given by you of the phenomenon have as a circumstance in common not only “has consumed water”, but also “has consumed an intoxicant”: gin, whiskey etc.

    The assumption here is that we don’t know that, never having tried them straight, and we are trying to find out why X-and-water intoxicates on that and only that basis. Mill’s Methods do lead us to water as the common factor in that case.

    the symbol of Wales (well, to us outsiders!), y ddraig goch

    In proper modern Brito-Romance, it’s ill dragun rhys, of course. Orthographic y and i are /i/ stressed and /ɪ/ unstressed, and u is [u] stressed and [ʊ] unstressed; stress remains final, as in Middle Welsh. ‘Fish’ is pisc. Sound changes from Vulgar Latin.

    (One fine day this sans serif font will drive me up the wall. Ill is the titlecased version of ill, but it looks like a @#$% Roman numeral.)

  61. Der er mange ege i den her skov = ‘There are many oaks in this forest’

    Really, den her?* In England, people talk about this ‘ere forest although it’s colloquial and old fashioned. I wonder what the connection is.

    *Norwegian Det er mange eiker i denne skogen. I don’t think you’d say ‘this here forest’ in Norwegian or I’d have noticed.

  62. David Eddyshaw says

    Which implies that CORPORA/corff must have been understood as a plural within the history of late Brythonic/early Welsh.

    Although it’s a distinct suffix in the phonological system of Middle Welsh, there is an Early Middle Welsh plural suffix -awr found in indigenous vocabulary (e.g. pennawr “heads”) which I imagine might have contributed to the interpreting of the -or of corffor as a synchronic plural suffix.

    Another -Vr plural is seen, of course, for brawd “brother”; this started out with the etymologically-expected plural broder, changed by analogy to brodyr, with a by-form brodoryon, which acquired the sense “natives” and a new analogical singular form brodor.

    BTW, I notice in Evans’ Middle Welsh grammar (which I should have looked at before) that sêr “stars” is actually an analogical reformation of the original sŷr, based on the singular form seren.

    Singular dragon notably turns up in Pendragon, though in that case I suppose you could declare it a genitive if you were determined to save the phenomena. There are other cases of fossilised oblique singulars (e.g. Caerdyf “Cardiff.”)

    Basically Welsh seems (as anhweol suggests) to have happily used whatever material came to hand for number marking, regardless of etymology, cheerfully remodelling singulars on the basis of the plural as well as plurals on the basis of the singular. Cymru am byth!

  63. David Eddyshaw says

    Looking up Pendragon, I see that I have fallen into the same error as Geoffrey of Monmouth (not a line I ever expected to write in this lifetime.) Not “dragon’s head” but (probably) “chief of warriors.”

  64. Lars (the original one) says

    Really, den her?

    Denne is book language by now in Danish, nobody says it spontaneously except in a few customary combinations (i.a., denne gang, denne vej, denne måde) but even in those den her isn’t markedly colloquial. My informant (mother, born 1935) agrees that this was the case even when she grew up, but I would guess denne has lost ground since then too..

    But this a very strong marker of spoken vs written language, even the archetypal misspelled janitorial notice will use denne. I wrote the sample sentences with den her because I wanted to judge how natural they sounded when spoken.

  65. Ooh, how interesting. Please thank your mother, Lars. Mine (b.1926) is also a useful informant about speech.

  66. In German, der / die / das hier is also typical for spoken / colloquial language.

  67. PlasticPaddy says

    English uses this here + noun, which is not possible in German (there the only correct form is this+noun+here) but present as above in Scandinavian. Is it (a) based on a French calque, i.e ceci, celle-ci or (b) a construction retained from when this and here both meant this (and this here was used for emphasis or as rhetoric) or (c) something else?

  68. I think combining a demonstrative pronoun with a local deictic adverb / particle is one of those things that can happen anytime, and then be grammaticalised, without any specific external trigger.

  69. David Marjanović says

    May I regale you with some subtractive morphonology? We can even keep the fishy example.

    I’ve mentioned Fisch, pl. Fische. In Upper German south of the zone with Inderior German Gonsonand Weagening, consonant length remains intact; and intervocalic and final sch always come from a consonant cluster (/sk/, later /sx/), so they’re long…

    …except that, in late MHG times, all word-final long consonants were shortened on the Bavarian side (I think not on the Alemannic side). This was not repeated after the next round of apocope, so that “fish” is now /fɪʃ/ sg., /fɪʃː/ pl..

    This short /ʃ/ found in the singular is a rare phoneme. It occurs in the few other singular nouns in -sch*, the 1sg and the singular imperative of the few verb roots in -sch (because those forms had lost their endings before the final-consonant shortening)**, in loans where it represents foreign /ʒ/ (mostly French***, one possibly Hungarian****), and in two vaguely sound-symbolic frequentative-turned-diminutive-or-something verbs: /vɪʃl̩n/, /pɪʃl̩n/ “to wee, to pee” – and I think that’s it.

    Thus, the plural is “unmarked” and the singular is “highly marked”, phonologically at least.

    (Unlike in High Alemannic, there’s no length contrast for consonants in word-initial position or when a nonsyllabic consonant follows.)

    This leads us to a singular remodeled after a plural remodeled after a singular. While the h in Schuh, pl. Schuhe, is pretty much ornamental, “shoe” in my dialect is /ʃʊɐ̯xː/ in singular and plural. The regular source of /xː/ on the Proto-West-Germanic levelb is intervocalic *k, while here we’re clearly looking at final *x, which regularly stays short: the singular and the plural would regularly end in short /x/. What must have happened is that the plural was adapted to the length alternation found in “fish” and many other nouns, and then the singular was adapted to another common pattern, which is identity of singular and plural – presumably because shoes occur in pairs more often than alone.

    * Tisch “table/desk” and Frosch “frog” (with umlaut: /froʃ/, /freʃː/) come to mind. Also, outside of nouns, falsch “wrong, false” because /l/ had apparently already been vocalized, but not Hirsch “deer” whose sch does not even come from *sk, but from a late assimilation */rs/ > /rʃ/. Oh, wait, but it does occur in a verb derived from Mörser “mortar”. Maybe deer are like shoes, see below.
    ** e.g. wischen “wipe”: /vɪʃːn̩/, 1sg and imp. sg. /vɪʃ/.
    *** BTW, -age has become productive enough to attach to at least two native roots.
    **** The Viennese nickname for Josef was /joʃɪ/; I suspect the Hungarian form József [joːʒæ̝f] is involved. The nickname suffix /ɪ/ is even shared with Hungarian, where Zoltán gives Zoli.

  70. David Marjanović says

    two vaguely sound-symbolic frequentative-turned-diminutive-or-something verbs:

    One more on the same pattern: /fɪʃl̩n/ “to stink of rotten fish”.

  71. Denne is book language by now in Danish

    (a) There are basically two Swedish alternatives corresponding to English
    ‘this’, ‘these’:
    1 Den här, det här, de här are common in spoken and informal written
    2 Denna/denne, detta, dessa occur mostly in written Swedish.

    – Philip Holmes and Ian Hinchliffe
    A Comprehensive Grammar
    3rd edition

  72. PlasticPaddy says

    This here goes back to (Late?) Middle English
    “This here rehercid blamyng and chalenging is defauti and vniust.”
    The Repressor of Over Much Blaming of the Clergy, Reginald Pecock
    Cambridge, University Library Kk.4.26 . (Pref. MS)

  73. Just a minor addendum: the point I made above on the earlier Welsh plurals “corfforoedd” and “corfforion” (consisting of “corffor” and a native Welsh suffix), implying the existence of an earlier “corffor” (CORPORA) as a plural of “corff” (CORPUS), is not, I think, original with me: I think I read it either in Joseph Loth’s “Les mots latins dans les langues brittoniques” or in Henry Lewis’ “Yr elfen ladin yn yr iaith gymaraeg”. My copies of both books now being in boxes, I’m afraid I cannot verify which of them made the point originally (if indeed either did).

  74. It’s in Lewis; cf. The Bulletin of the Board of Celtic Studies, Volumes 11-13, p. 138: “Buwyd yn eu cyfrif fel math o luosogion dwbl y gair corff: *corffor (< corpora) + –yon, –oedd, gw. G., a Lewis, Yr Elfen Ladin, 35, d.g. ‘corff’.”

  75. John Cowan says

    This here rehercid blamyng and chalenging is defauti and vniust

    Alas, I think this means “This here-rehearsed blaming and challenging …”, that is, this blaming and challenging that I have laid out in detail here.

  76. Yeah, I agree.

  77. PlasticPaddy says

    Ok. How about (found on MED site by searching “this here” under Quotations)
    God forbede þat ony Cristene man understonde þat þis here synsynge and criynge þat men usen now be þe beste servyce of a prest.
    De Stipendiis Ministrorum (Wycliffite tract)
    And whanne thei coude not justifie this here dede and ryot, thanne thei voysed that it was the proper and several ground to the hous of Bury.
    (1449) Let.Garneys in SIANH 4 (Add 19112) “Rose Hall, Beccles,” Proceedings of the Suffolk Institute of Archaeology and Natural History 4 (1874).95-96.

  78. John Cowan says

    They work for me!

  79. David Marjanović says

    I wonder if this here X in these examples meant this their X “this X of theirs” – which occurs in archaizing German: this here dede would translate as diese ihre Tat.

  80. John Cowan says

    Archaizing English, too. “For the Beauty of the Earth”, an originally Anglo-Catholic hymn written in 1864, has a refrain of “Christ, our God, to Thee we raise / This our sacrifice of praise”, meaning the Eucharist. There are many variants, including the one I learned, which is not even specifically Christian: “Lord of all, to thee we raise / This our hymn of grateful praise.” So the verb raise, originally ‘lift up the chalice’, has been reinterpreted as ‘raise our voices’.

    YouTube version, about half the proper speed from my perspective.

  81. PlasticPaddy says

    David woud seem to have a point if hir=their can be spelt here, but this fits better for the quote from the letter. The quote from the tract seems to be Northern and I am not sure they used hir up North. Also “this [their]… that men usen'” seems to me a bit clumsy. By the way I thought synsynge was an error for syngynge but there is an ON verb syna “to show” which would fit in the sense of “exhibition”.

  82. David Marjanović says

    Also “this [their]… that men usen’” seems to me a bit clumsy.



    I wondered about sin-singing, i.e. it’s sinful.

  83. PlasticPaddy says

    The imperative phrase “syn sing” is used in Scots dialect poetry but I have never come across it in speech and could only provide a guess as to its meaning or its relevance to the Lollards or their opponents.

  84. John Cowan says

    In proper modern Brito-Romance,

    as I forgot to point out, the number is often carried solely on the definite article in both speech and writing: ill abitant, llo abitant. However the feminine article triggers soft mutation and the plural articles spirant mutation, so we have cases like ill camp, llo champ and lla gas, lla chas ‘the house/houses’. But there is no trace of the Latin plural endings.

    There are a handful of Romanian/Italian-style neuters: ill bwsc, lla fwsc and likewise bwyrdd ‘hut’, ew ‘egg’, ws ‘bone’, rham ‘branch’. There is another handful of frozen duals that are now plural: ill breich, lla ddewfreich ‘arms, pair of arms’ and likewise cil /sil/ ‘eyebrow’, cwrn ‘horn’, genygl /ˈdʒenɪgɪl/ ‘knee’ (with echoic vowel), llavr /ˈɬavər/ ‘lip’.

  85. God forbede þat ony Cristene man understonde þat þis here synsynge

    A footnote on the last word in this edition of Wyclif says “incensing.” The OED s.v. cense, v.¹ doesn’t have the spelling “synse,” but it hasn’t been updated since 1889.

  86. >in proper modern Brito-Romance

    Is this a conlang you’re creating? Googling offers me various Britos who have authored novels or been involved in romantic comedies.

  87. Yeah, I think it’s JC’s personal thing.

  88. John Cowan says

    By no means mine. Brithenig was invented by Intheologus, mundanely Andrew Smith, of New Zealand, and is now real enough to have a ISO 639-3 language tag ‘bzt’, a WP article, and various entries in online dictionaries including Wiktionary. Here’s a nice little poem by R.S. Thomas from the chrestomathy:

    Perch ddef-eo yscrifer sig?
    Eo su kemruis, gwiddeth?
    Yn Comro pyr,
    Torb i llo few wuin.

    Sa’i riglafaf mui,
    dugef-mi afont mellt,
    Rhen ddyrid;
    Sulfent ill yn dawn,

    Eo mhod rhen parlar mew
    Llinghedig probr – Iesu,
    Llo pharol fon thud;
    E affor llo…

    My only contribution to the language, I think, is the name of the village with the longest name in all Kemr: Pluifairllagunblancoryllentiostillrhebiddgurypluitysiliocafurnrys.

  89. Ah, thanks. I only knew of it from your comments, so I made the natural assumption.

  90. den her

    Estonian also has see ‘this’ + too ‘that’ that can combine with siin ‘here’ /seal ‘there’.

    More on this:

  91. I followed the link to find the etymology of the American use of creek. They give “up the creek without a paddle” as a more intense, later version of “up the creek.” That’s backwards. Up the creek is a shortened form of up the creek without a paddle,” meaning got yourself into something that you’ll have a harder time getting out of. On its own, up the creek isn’t a meaningful expression.

    They give up the creek as going back to 1941, but I get a “without a paddle” reference from the Suwannee Review in 1892. Split by a few words so it may not have shown up in the search they used, but certainly with the phrase in mind. And a 1915 “up Salt Creek without a paddle,” which is euphemistic. By the late teens the google hits for without a paddle are everywhere.

  92. If you’re up the creek you can just float back down, without a paddle. I guess one just shouldn’t be so literal about it as long as the expression sounds good.

  93. I agree, it doesn’t make sense.

    Maybe it originally meant that you are supposed to go on a boat up the creek, without a paddle.

  94. On the other hand, if the creek in question was something like this

  95. In Green’s dictionary, Up Shit Creek predates Up the Creek by a lot (1868 vs. 1925), so the latter must be a euphemism for the former, and the paddle is more poetic than practical. Class 3 whitewater (mutatis mutandis) shit creeks are intentionally beyond anyone’s imagination.

  96. Ryan said: I followed the link to find the etymology of the American use of creek.

    That link goes to Etymonline, which just summarizes the old OED, as quoted above by AJP Crown. That’s outdated; the sub-entry “up the creek” was written for the 1972 Supplement, so you can’t expect the same completeness of citations you can get today.

    As Y notes, “up the creek” is euphemistic for “up shit creek” (which wasn’t entered in the OED until the 1986 Supplement; I guess that’s why they failed to cross-reference it from “up the creek”). Up shit creek is self-evidently bad, and floating down wouldn’t be any improvement, but a paddle might give you some control or even help you get to the bank and out of the creek. Green’s gives even more variants—in leaky gumboots, in a barbed-wire canoe—which demonstrate that “shit creek” is the nucleus and all the rest is elaboration.

  97. I get a “without a paddle” reference from the Suwannee Review in 1892.

    1892 is the date the periodical started, not the date of publication of the issue. Google often does that with periodicals. The snippet is:

    An entire generation of the English intelligentsia was up the creek. But not, fortunately, without a paddle. The paddle was Dylan Thomas.

    So definitely not 19th century. This appears to be vol. 66, dated 1958.

  98. Reminded me the joke about two strange animals “grebiblya”, and “grebublya” usually found in a [three-person] kayak.
    grebi is “row!”
    grebu is “I am rowing”

    blya is “whore” (here cliticized)

  99. Well, ‘whore’ is the literal meaning, but that doesn’t help much in this context. We discussed blya(d’) at LH in 2004 (see also).

  100. Lars Mathiesen says

    A very strong marker of spoken vs written language — I never reported this but a year later I spotted a sign on a door saying something like Den her dør må ikke efterlades åben. So denne is even losing the domain of signage; recency illusion and all that, but I do think it is new.

Speak Your Mind