BREV TIL: Stephen Anderson FRA: Eli Fischer-Jørgensen (1983-10-31)

3l/lo 83

I .


Deur °teve, X'hank you ver s ocD or ''end in,. me various diopters of your b> and r^c fitly the cbop+e1’" or ,glo- era tier, and exe« e me for not bovin,? reacted to it. reason is that I have been fcerrib- ly bury io - Lhr lost tw y an..- • bu ;y n try T i ""e. tvben I ret ■ ed I thought that X would hove plenty o tire, and I accepted all nvitat Ions to contribu e to .featsc hift-v lu , 3 tripe? ia etc. i ith the result that 1 have hardly tied a in' tiu ■ e free. - This summer- -was particularly hard, but X finally decided to go on holiday for a month at the end of epteber. I got y ur chap er or plosso <> at I as the day before leaving1, o X. put it into the -suitcase, but X had to loo'^ various thin s up, so I was not able to answer tmtil now, I think you have done a very good job. It is very difficult to write about gloasenatics, but your account is very clear, and. your criticism convincing, I have only a few remarks, - ( ■'■here i a hum er of misprints, but X think you have found their, yourself ),


■nve, in fact ,never been so

€> development

As foi’ the first '•■or hi torical part on of glossemat* cs -X can only agree. As far ai? I can see you hove based if -rt ly n my ob* inari < etc. - Th re 1 r nl y one point x-f ore X disagree completely - ^ p. 217? 1. you mention rue t- ret or with nienae Was . Verne and Jespei . X ma ly do not belong* inthat c.npnnyl I have ot made nuy' centr > but ions to ling is :ics at all* hove written sore critical papers on phonology, which are rather clear,but nog original, and X hsve contribu ed to the history of phonolo- t5,at n'l# - You c Id mention lots oi' Danish linguists

To re. below, hoi ger ■‘■eder' on

fry, who who have given much more important contributions t Sand- fed t r, al, Hide ch en, |o >.eby, Lech, gerod, iXisolabl. plea e, leave me out l(Eut it was nice of you). - Ey the waylljelmsliv'rf books have been translated Into various la*if'ua es in Sot ern Europe during the latter years, ( ranch, It liars. Spanish, *vu anian etc.), jo t ere is , in fact i an inter- st in his t eori s, som« what more than ajpears from xvbat you say on p. 216,

P;~21 X am more inclined to a ree wit Diderich en. Of course Rask knexv that Innguage changes, ^ut he did not distingwii clearly between typological end enetic relat ns ip. p.23o ff y u use the term realization, where Hjel«slev^said manifestation, X thin1: y • • u rko- 1«1 ..ticir t mani f o»t at ion. t/jelme lev very caef ully avoided the tern realizati>n in this connec- t ion.

ini' you say fat Hidal 1 he a succession of positions in Sndon, Arge tine etc. x a not urc that lie hod any position in ■L/ondon. I think he only stay: d there for a sl orttirae, -1- can




a sic Betsy - p,239 begin- in of' lost paragraph, q-vite correct. Accord ' n to the Résumé? ta semes are the last units you arrive at in the partic lar analysis. But y u may the1 c ntinee mnkin an univ r a analysis, dissolving the taxenes into prime factors , ca ’ led glossernes ( w' 1 ch are designated by the Greek letter A o4_ to. Exr> ssi ?n glossernes are calif


I do not think t i i


cenemes and content g"ossetnea are called pierernes,inso far as they are constituents, not pro codecs*? a. - ue has used the terms cenernes and pierer os somewhat moreloesely earlier, but still I t ink the be t tbiup would be to omit He n es "coner stxc" and pi. r erratic" units more gener« ly for express-*' n a -d content unit •, - ( probably "unit Is not the right go e -tic ter- either ) it has a more spec:if?c meaning and he talks about pleres -t cs nd cenemat i os -- taut -- I know ■** have usd cenemes and t;7 or ernes In my phi ology book in two pie Cx,f s acc rd'ng to 1 > ea li r and later use, which is confusing’, * are not r - i n and con . ant . •

only for con titr nts. t q terms.

in any a e v t ‘xeiji.?3 are a ptoxmtely of' the p. 2ko 'T ' 1 cene~ es^

sa*”e s * ze as ph, . lu; lev divided the expression elemei ee >rrf c lanse •• of rules : distributional

al§?r T'

^ut this is not glo.

clustering, alter at rs and implication, matics. This . s t- e d:«vi:: ’in he made in the paper from 1935« In 1 i 'os e atc de cr:tpti sn he A’«tint grouped the ex ression taxemes into categories accord u to di. tributional clustering, - JTixhixxKHX»rr'-pr pul sfrJtxwtritn^atxhe .;d rxx .edxevgv . t rie c .mxv.nm.niiB intcriw f. 11 a ± x x x tr. *:: f i wa i s K d x fex X fc x nn at * x hh xsi x ± x in e « x h g h x » x i w x * k o s e xxxuk. 'SchHXJBKHxtitaHMtrtxx. Th " ther,niter" are d v ded into vowels and conson nts as selected and select?n , ( plus bot: and- and neither -rior.)JLhc consonon.s are further divided txtx according to t ■ ir possibility of be-'ng initial and final • c !.o e , to the vowel or not ( see his e on of the French systen;/ • but this* is the "popular" version. In the strict theory it was a quest 1 or of deter i ati n el c. pres positions of^^liffe kinds . °yncreti fur« only en ers in the os• eu o-t .< a 1 .yV'fs u^ter mining n,- ©n<s are xb4sk intensive and which are extensive.


p. 244 ,1. 11 rom below r b: h where f ha 7 b nd p do not contrast i at least after* s> orant c nsonant)" I do not under- stand t i>. ansih pt; are bdg do not c ntrøst - nelly o redial b fo e ..chwa, :i . e. what H j. would c^l in c.yl loble final politiet , ’-'hi : i true jjszpz i reapect ve of the preceding taxeme. The i ,,e^;^ay \>e a contrast in a con ervative norm a - ter a som rant conrrawt p’u.i schwa, e.g. ve den v< rten, pu pe , bor. le feste mainly in the length of the sonorant , and more voicing in the consonant - kul so just af r sonorant there may be a distinction ,but only i ax conservative norm. - ''hat were you thinking of ?

p. 246

, -9 Here you seem to h ve written "form" in stead of "expression’,' as opposed to content

p.247, 1.12 iron belcw voice;!, read voiceless , but better aspirated st ops, since t' ere i no contrastof vetoing in Danish stops, 4 eXefore you should not talk either o a rule that would "devoice " stops in the vicinity of h» - The justificati n of an analysis of banish asoiratel st s into stop plus h seems ratiierobviou to me Since i'e only difference between pile and bile is the upirat; a, i± and since h is cund as an Independen j consonant in Danish, i* i - very te pt'rr to analyse p n b + h . { this is what * s also often done in simple plien tic transcript! i of Danish: oils written bhi si , “*ut ,, when iAj elms levs write? hb hehas 1 ft substa tia1 motivations.

p. 248 ,5 from below. A'-a in I do not like the ter "Voiced .top plus h, " Hie American habit of ca ling ptk voiceless and bdg vo oicel pi and voiced as la' els for ptk and bdg, is rather confusing -«nd make it more difficult to unders tand th analysis. ~ in loanwords

wed, i e. u • •* n r

Mo .over hjelmslev add ced cases of alternations , e . g. lak lakere, il.-. ilagh?


j Iro 1 v do 3 not treat stress as a ela i:i nal property be w on sy i lables" Thi« 1. find soi ewhat mi ’leading accents Kxixdsf nKd'byx as exronetPs ar def n<d by directi

ft IX


W-; fr

m -



tress iresupposef j« 1‘v-lev only di tin -v ished two de^rsif s of c Tat oh t.i-j ■ ht tn e pl^ce at d.‘ ;fe ont level«

a nd tin ns »elati nal elment . ^tr ncstres-. and 3tress, tyut t i

fr v



p. 2 =58

lo trom be ow, X & not q. 1te ■ re wl ■ t y u mean by ti>© en ence •’ Aetna 1 >y , sirn-e d,je tredve -itv ara to count




-- X il s’»ov tbi chapter to xVi cteel. Perhaps ho pay "••o ve < me o iiient s. ^ hope to h. ve time to read the rest of you r manu- script soon, at t e Into•t in January.

Ilest regards







- .. -- ■ V ■ •- ••

-- .. • - --^p- ---


‘ -





■ r


■ /

AC-: '•iV '


■» A




.. >.