Sauwue’s Theory of Language Sowsune yas enhappy with Tingeisties as he knew ie Irecante he thought thar his predecesors haa failed to think sevioweiy ur yierceptively about shat they were doing. Linguistics, he werate,! never attempted to: determine {hi suture ef the object tt was studying, and without hi ‘slementary operation a eciener cannot develop aa =ppro- priate method? (arse 33 Cmie, 18) ‘This operation is all the more necersury because bum language is an extremely complex aad heterogeneous phenomenon. Even a single spceeh set involves an extra fordinary range of factors and) could be considered fom: ‘many different, even eonilicting paints of view. One could stuily the way sounds are produced by she mouth, vocal cords, and tongue, one could investigate the sound’ waves whicit mre emited and the way they affect the hearing mechanism. One gould consider the signifying intention 9° the speaker, the aspects of the world 10 which die utterance refer, the famnediate cixeumstances of the communicative Context which might have led him 10 produce m pasticular series of noises. One might iry to analyse the conventions Which enable spesker and listeners to understand one nother, soeking ont the grammatical and semantic rates which they must have asimilated if they are to communicate in this way. Or again, one could trace the bistory of the leoguage which makes available these particular iors at this time, Confrosted with all Giese phenomena and these different perspecives from which ome might approach them, the linguist muse ave hirmnelf what he s trying to describe, What in pacricular is he looking ac? Whac i he louking for? What, inshore aaguage? Saussie's answer (this question is unexcoprionabile but 18 Senses Theor of Language oremely important, since ft serves to divest attention 19 eseniais: Language ts & ayatem of signs. Noite count a5 hhoygusge only sehen shey serve to express or communicate ss; otherwhse they are just noise. And to enmmnnicate ines they must be part oF 8 system of conventions, part ofa system of signs, The sign isthe union of a form which signk lis, which Sawsaure calls the nigra or signifier, and an len Sgnified, the sign signined. Though sve may seal Ot Sentfer and signified as if they ware separate entities, they exist only as components of the aga. ‘The sign is the enteal inet of language, and checefore in tryin to separate what is essential Irom what is seenndary or incicental we must start irom the mature of the sign ite, ‘The fest principle of Saussure’s theory of language com cerns the ewentil quality of the sign. ‘The linguistic sigu i arbuwary. A particular colubination of signifier and signie Sed i an arbitrary emicy. This a central Lac of language and linguissie method. Noone’, he wrices, ‘the prinespe othe arinary sete ofthe sm, bu it ‘lve ease fo disover a truth han ia assga i ight place "The above principle docsinats the whole of inguste smalyais of language Ils Gorsnquonets are innumerable, chong hey 86 fot alls Wve, equally widen stssighe away. [te sher Trae tours that one dicovers them, and wits them the fandsientd Imporcance of this paaciple (Cars, 8; Cas, 100), ‘What docs Saussure man by the arbitrary nazure of the sign? In one sense the answer is quite simple, There is 9 ingjural of inevitable link hetween the signifier ane the fighilieg. Since [speak Pnglisn I snay ase the signslice represented by do to talk about an animal of a partenlar species, but this sequence of sounds is no better suited to ‘hoe purpose than atother sequence. Lad, let bop would scree equally well if they were accepted by members of ry apeech community, There is mo intrinsic ceason why 19Seusise shades of a single colour are in Russian ti distines primary colours. Each language articiles or enganires the world diferently. Langinges do. not simply pane esistig. eategories, they artculaze their ow. Morciver, if lnngnage were a set of names applied to independertly-existng concepts; then in’ the historical evolution of a language the concepts should eemain sable Signifies coald evolve; the particular sequence of sounds rsscciated with a given concept might be modified; and a given sequence of sounde eauld even be attached to a Jilferent concept. Occarionally, oF cour, a new sign would have to be intradkced for a new concept which had been, produced by changes in the world. But the concept chem. Selves, as language-independent eabies, would not be subject (linguistic evolution, Tn inet, diough, she history of languages ie fll of examples ot concept shitting, changing their boundaries. The English ‘word cath, for esample, at one point meant property in eacral, chen gradually came to Ue revirieted to fourfooted properey only (a new category), and finally atiained its rmodern sense of doxnasiivated bavines. Or again, a ily person was ence happy, blessed, and pious. Grachally this particular coneept alcered: tae old cansept of sllines? transformed itll, and by the beginsing of the sixteenth century 2 ally perton was innocent, helpless, even deserving ofpity. The alterarion althe concept comtimaed until event allya ally person was simple, Scolih, peshaps even stupic IFlanguage were a nomenclanire me should be obliged to say that there exis a number of sstinet concep and that the signifier si was attached! fet to one ad then to nether. But clearly this 3s not what happened: the cone opt attacned tothe signifier sly was continually shiftog i Tpotndavies, gradually changing its semantic. shape, articulating the world in diferent ways from one period t0 the mext. And, incidentally, the signifier also evolved, Lndergoing a modifeation ofits central vowel, ‘What isthe significance of this? What does ic have to do with the aebitrary aature of the sign? Language is nota . Senses Thecey of Lenguage mienelntare aad therefore ite signifeds are not pre- ‘ung concepts but changeable and consingens concepts ‘vlucl vary fem one state of 9 language fo another. And ince the relation between signifier and signified is arbitrary, Since thse ine necestnry ranean far one concept rather Ua nother lo be attacked to a given signifer, there is Iherefore no defining property which the concepe rust recain in order to count a the signified of that signifier The sigaiGied associated with x sigafice ean take any oem; here i no essential core of meaning that it rust retain 5a foedes (9 count a6 the proper signlied fer cat signifer, The fact tha dhe celation bewveen siguiier and signted is arbivary means, then, hat siace dhere are no xed universal oneepss or fixed wniversal signifier, the signified ssl! fg arbierary, ond se & the signifier, We then must ad, a Scussure docs, what Genes a signifier ora vigified, and the anmwer leads us to a very important priaciple: basi ‘ignifiee and sigaified are purely relational or dilTerenval tntities, Booauve they are arbitrary they are relauonal. This isa principle which requires explanation. THE NATORE OF LINoUISTIG UES Saussure ataches great inportanoe~ mores than t would tppcsr fom the published Chr to the tet ht language Gruden the illite of he acbirary nature of the og A language dcr ae snp en ariray nas wee of eepetnty xing concen It Sop a Sluary relation betwoen sigueer of ina chocong Or the ene band, and sped of ts oun choos one oie Nev oly docs excl Innquge produce wale sc Gf agnites, ardcalsting and didi he eoaunuuon af found eo Snare way rach nuginge produce: 3 {Eitreon soc of suited © hos a tine and dn friar” way of orguizing the world leo concep 9 T's sbviow that the sound sequences of fle and 23Seowsnge oe ol these gifier rather than aoe howl Be fi with the concept of a ‘dog’. fre re no caceptons fo thi Disc priciple? Ceviinly There are two ways in which guise sigs mony, he module, Ghat uta say, made less astray Fis, there fe cates of onontatopai, wire he sotnd ofthe mpi geome in sore way taanetic or imitative, as in the Bris owns or era (eb French auad, Gerien sce, Traian fads). But there ae fe sich cas and Ure fat chaz dently thease aseparste clas and ypcial eae only tmphasres mace strongly tha oninary sign are abiteay Howeven within 2 parser Lsngange signe ay bs tially motivated im a eiBerent way. The machine on Shieh Tam tng aleg a peur There iene Yeason vy shold nat he called agra or a Bir, Bt ‘Sinn Engish pre motivated Because the meanings Sr the yo sound sequances which compos sign, ope foe soe, noe eles t0 Simi, tthe Noten of 6 “Sypenrite. We might call this “ccondary notation ‘Notes, for empl, thet only tn English fe the relation bensecn srndsequenoe and concept mesivat. If the French were to ase te sme form to speak ofthis chin, that would be a wholly arbitrary sgh, since the penny consitvent, vier iy not mgm fn the French language Kinveove, for snare, a8 we shal ce Ines, the proses: combi ie anc rite erestew ew mtd is tandamentally salar 10 the yng in wich oe eorbe words to form Pheer (wheve tening related tthe Combined mneaingy oF indvadunl wort). We can say theveove, dt all languages have av thei basic elements arliuary sig. They thet have vasous proces for om ining these igus, bu thas does not alter the cool aru of language and tg elemensary conseace “The sign ie arbiray in thar there fe no inerinie link between sister and signified, ‘This i Pow Saussure’ principle ie sally Tacepretd, but in thie form ie 8 2p and seonton nara degite eng eg. Wop Seusie's Theory of Lonenaee ly. tndiional notion, a eather obions fact ohnnt uate Lmenpted sth Tepes ony Ht eoos nl DE She manent consequence wish, seeording othe cose, Sasute repro chimed fer ie he iitiurechiea place ofthis wath ren the very som Te ty te by lite at ene recogetes how tansy erent ‘hots are but ramilestions, hidden consequences of ths rth gles, 199) There is more tothe arbitrary native ‘ithe ign than the arbitary relation between signer 288 Slguifecs We niu push futher. From wat {have ide fr sib siguifier and signed, one might be erapted to think o! langage ae Nomen Clases series afaumes aricanly selected aad attached tora ol of ebjecs or concept It iby Saute sash all oo tncy to think of language asa set ol ames adc make the Sic story of Asa emning te Deas am acount o the Srey nae of language, Tone says tak Ue concep “eo [rendered ov expresed bye i Engi on in Pre {nd fin Corman, one implies that eae argues a arbitrary name for a concept which exits peior to and indepententy of ny ln Wiangunge wore simply s nomenclature fer a set of mivens concept to Be easy co tussle rom one Inngonge o ante. Ce oul spy repiace he Frcs fume ora concept withthe Baga ame lunge re The thi he tos of amin a now language wile abo be tc cote than ts But syone who han ered ether Giese take nes soquired, aig, vase somone of ect rot thet Tangeages ive tot nomendanures thor the Eoncept or stieds of one ianginge tay ditfer rately fron those ofanother, The Freneh nner’ doe a0" gD “heatly to Englahy one mst choose between to ke” ad “lo lwe! ‘Demucner! includes ma single ten the Engi figuitccs of comving oll” ad “aceclentng” ng" ‘how saver the aera of uo French signified, onal” fd hs Tne Engin conorps a eaekee nn on a ‘per hace uo sea counterprve in Facil Or again, what Bogus calle “hgh blac? se! ck Blue and estSovesrs hitw ave cgnifiere of French but age of Engl whereas coer and steam ace English 21 20% French, Les obviously ‘but miere significantly, the organization uf the conoepeaal plaus ie ake diffrent in Eoglich ord French. The signified ‘siucs ie opposed to ‘sercam’ cole in corms of sie, vhereas ‘a "Beave” differs rom a ‘rvitre" nor because itis necessarily Targor bur because flaws inta che ses, while ‘civitre” dose nt In sort 'fleuve” and 'rivitr’ are no: signified or cone opis of English, They represent 2 diflereat acticlation of the eoacephial plane, The fhev chat these two engages onsrate pesfectly wel ifivent conceptoal arieslations of dieriesions indicates that these divisions are ret natural, snevieable, or Ineceetary, bt, im an inypavtant see, avbitetey. Obviteay i¢'s lnepaetant that a language hes ways of calking sbomt Mowing Bodies of water, but it enn teake ice coneeptial distinctions an this ares in any ot'a wide variety of ways {size illness of flow, strsightness or sinuosity, direction of Aye’ depts, navigabilisy, ec.) Not only cam # langue arbitrary choowe te signifies ir can divide up & spesieus oF ennceptsal possibilities in any way likes. Movetves, and here we Couns to au important point, the face that thai conceps nr signified are arbitrary divisions ol a continuum meam that they are not autonomous cotities, cack of which is defined oy some kind of esence, Tacy are members of a system and are defined by their relations to she other members of that syatem, If Tam 10 explain fo someone the meaning of szear I muse tll him fsbou! the eillerenee belireen a seca anda river & ete ec. Aud sunlarly, | cannot explain uic Peck concept of a ‘viviére’ “without deseibing she Uisuxetion setwoeu ‘islére™ and ‘fleuve” ot the ome hand anil sviére and naiseau’ on the olen Colour terms are @ parsicalarly yued example of this haractevisie of the signe Suppose we with co ech a, Joveigner about colours ip Bagish. Ler us suppose also hat hig a riuher dow learner fom a non-Buyoaean culture, 30 that we sus work vut an ellieene ceaching stracegy. TL 24 _ ~~ Seuscr's Thay of Language Is occur to us that the best way to praceed would be 29 Fite one colour at a time: to begin, far exaals, with Tinowi ad ot go om 2 another colour neil we were svrtain that he hist mastered brown So we begin hy show= ny hm browa objoets and welling him that they are brows. Since we want to be thorough, we have assembled a ‘ollectin ofa hundred brown objects of various Iinds. And ‘hen, aller having bored him and ourselves for several hours, tre tale hn into anotler eoom and te (est hus knowledge of ak hin co pick out all the boown objects, He set an work but serms eo be having diffeuly deciding wha w seleer, 19 in despair we decice we haven't been thorough enongh and propede to scart agua the nest day with five Inundred brown objects. Fortunately, snot of us qwoulel not adopt this desperste solution and sould recognize what had gone wrong, How- fever many brown objects we may show fea, ox pupil Wil hot Know che ateaning of brim, ans will not be able pss cou test, lil ve ave taught bio te disingaich | between brixen and red browa and taa, brown and greys boron and yellow, Drone amd Black. Te is only wes Ae 28 prospre the relation between Brown ane other 69/045 that he will agin to uncerstanch what brown is. And she eaon fo this that Srows. ts not an independent fined by sone essential properties Bue one teem ima fysiem of colour term, defined hy its selabent with the other teri which elect, Indeed, this painful teaching expericner would bring ws to understand that because the sgn is acbiteary Js he sult ofdividing # coudouum ix ways pecul Tnguage to whi 3 beloags, we cauunoe treat dhe sign a8 ‘an autonomous entity bur iaust see it as part of a system. Tis noi Seat that in order © know che meaning of beam fone suet understand ve ten, org. blac, oie. Rader, one ‘could sayy thas che signiiede of colour terrae ave nating int the produc: of rele of a rystem of distinctions. Egeh Thnguage, im dividing she specwum and diiuyishing atngorles which suis eeoury, produces a dilloren eystemSeucore ‘ol signifiede: nits whose valtie depends on their relation ‘with one another. Ax Savers says, gencralieng the point in all ease, then, we discover not aber given in scence Sut tates erating from the system, When we say tha these Pls smespane to concep ot we sndestond thot theke anceps Ate purely eifrendal, not posidvaly coined by chelr concant bur Dauively defined Oy aheir relations with other vere ot the System, Their inet precise characteristic fs thar they are wht the adiers are tot (Cine, 173 Gets, 10). Brown is ehat is mot red, black, grey, yellow, ete, and the same holds lor each af the other sigan, Thus x # major thougi paradoxiea! consequence of the arbigany natice ol the sgn, and we sal rear to it hor ‘But perhaps the easis! way co grasp this notion ofthe purely yelavonal gature of hinguistic unis is to approach it irom another angle Consider the problem of identity in linguistics: he ‘question 0! when two uterances or postions of an utterance count at excinples ofa single fiagasic unit. Suppose some fone tells me, “{ boughe a bed teday”, and ( ceply. “What fore of bed” Whist do ive msan wien we say that the same gu has been employed wsice ia this brie! coaversation? ‘What is the bass on which we can claim thar two examples or instances othe same linguistic wnic have appeaced in our dialogue? Nove that we have already begged the question ‘tranceribing 2 portion of the noagge that each of ue rade ‘5 bea, Tn fact, the actual noiaes which were produced sill have been measurship different — diferent from 2 purely physion! and acouie point of view. Voices varys after a very few words we can recognize a friend's voice on the telephone necanse the actnal physical signals she emits are dlflrens from those of ou other aequuaintanees ‘My interloeator and 7 prosiuces diferent noises, yer we want to say that we have produced the same signifier, wed the same siga, The signifier, then, is nat the same thing as the noise that viher she ur procuced, isan abstract unit ofsome kind, not io be confused with the actual sequence of sounds. Bur wha: soct of unit is i? OF what das the unit , Scusrne's Theory of Langsags ‘onnist? We sight approach this question by sag how Fir ihe actual noises produerd enld vary and aul owt vs versions of the same signifier, This, of courte, i Similar Jo-shiequestion we impheittyasked enrles bourne ignited: Inu far ean a coleme vasy and stil couse as brown? And the answer for the signifier is very similar 10 the ansvee for the signified. The noises made can vary considerably (there ietao enential properey whieh dhey must posses) s0 long as they de not become confused with thaw of contrasting signiiers We have considerable Tatiuade in the say we ulier ded, so long as what we say is net confused with Sad, dy Sc ba Day lad dh, fy a, lech ra sa, sds duc, bel, bt Tn other Words, itis the distinctions which are important, and itis for chis veason that linguistic units have a purely relational idencity. The principle isnot easy to grasp, bat ‘Sauseaze offers concrete analogy. We are willing to gra Wat in an imporianc sense the 8:25 Genevarto-Pacis Express isthe saune train each day, even though the coaches, Tecostotive, and persoansl change from one day to ie nex. ‘What gives the cain its identi i its place in the pater of vwalns, as indicated by the thnetable. And note that this relasignal idenity ie indeed the deiermining factor: che ‘rain remains the sume qrain even (it leaves hall an hour Jate, Indeed, it night always leave lace without ceasing to be the 8:25 Genev-io-Paris Express. What is important is that iz be Giinguied frum, e8y, the 10:25 Genoraso- Paris Expres, the 8:0 Genev-to-Dijon local, exe ‘Another analogy which Saussure uses to flustraze the rotion of relational identity is the comparison between Janguage and chess. The basic units of eness aro obviously king, queen, raok, kaighs, bishop, and payin, ‘The acival physical shape of the pieces and the material from whieh They are made fs oF no inigortanee, The king may be of any size and shape, as long as there are ways of ciingoishing ie from other pieces. Morcover, the twa rook cee ot be ff identien! site and shape, #0 long as they can be distine [puisned ftom other pieces Thus, ut Sausaure points out, ifSaesere a plece Ss let om a chess st we can replage tt wih any fvher sort of abject, provedel alvage that this object wil Sot be confine wth the oboe reprsentig pee of = Giheviar value Coney 1101 Goa, tarde The accel Shrslal propertin of pieces are of 80 Importance, so lng. ovhave br cierenee of sore Kind amy Ion wll 0 = Temmecn pest whieh have a eifleren vale. TThavone cam say that che unit ofthe game of ches hace notarial ens Were are no. phycal proper Teceisy to a hing, cies Beatty Se whobiy n neton oF {Eifrenco within a sytem, Trane ow aly the analy Colanguage weshal bina portion to understand Sasasures paradenival claim that in the sytem of « language "here Ere only diflerence, with no positive terme (Cnn, 1205 Goes, 166), Normally when we thik of dllecences we frosuppere two things wich dex but Sausmure’s point D'thet signer and sgulied are nov chings in this sease. _Jostas we can't sy anything about what a paw must lok Tt, exept that (vl be deren Grom igh oo, ey wp thesigitie whos we represent ae Be eto" defined by My pariovnr noes wed neering s. Not only do che clon noes cifer from ane case (9 another, but Eelish foul he arranged +0 that noes ow Wed 0 exprens the Seer pl eee for the sgnier fe and vice versa, TF these changes weve made the ons ol the language would be exptesed diferent, but they would stil be fendameataly {he sane ats (the sme aforenees remain, both on the Jevel ofthe sghterand.on the level af the signed, and ‘Fevanguage sould sil be Pragith Indes, Binh woe Fenaine ina imporlant sence, te same language i the fnitsof the signifier were never eaprened in sound but only in vista symbols oF some kid. In saying this we are obviously making @ ditnction between unis af he lngulse sates on dhe ote hand sod thew actual physical manifenations o¢ realizauons oa the Sion Pelore daning es very imporsit dstnesion ‘ente detnl it may te seh to revapialace the Hite of veasoning that led us to fe. We began by noding thar there 8 ~~ Souscs’s Thoory of Language wos nes antural Tink between signifier and signified, and thos, tying to explain the arbitrary nature of the linguistic inane sans that both signiter and signified were arb teary * Signs dsligaiationtot a eontintiam {a sand spectra ‘Uke owe hast andl a eoneeptoal fele on the other}. This Tiel fo infer that Lod signifier and sigeitied mst be [ote in terms of their relations with ether signifiers and wuuiieds, and Uhus we reaches te conclusion that if we are ti define the units of a language we most divingush Deqween these purely relational and abstract nits and their plinsical realizations. The aewaal sounds we produce in Zjwating are noc in themselves unis of dhe Lagu systema, tor is she phytical colour which we deskuate x calling a book “brown” dhe cause Wigg as the linguistic unit (the ssepified or concept) ‘brown’. In both eases, and this is a point on whic Saussure righdy insist, the linguistic unit ie form raiher than subsance, defined by the relatiuns wohieh set i off from! other units, Hore, ia the aistinetion hetsen the linguistic system an ite aviusl manifesavans, we have veached the emicial oppress Soom betsess forge ain pte. La tvgne ve the systerd of & language, te lange as aystem of ems, whereas parle the speceh acts whicl are male pemisle boy dhe Lusguage, Li lange is what the ieividal assimilates when be learos a language, 4 set of forms or ‘board cepesited by the practice of speach ut speakers who belong io be ime commonity, a grammatical system whieh, to all lnuents aad pwpeses, exats in die qund of each speaker Gener, 1-043 Ceri, 30), ‘IC's the social product whore exisionce permite the ‘individual 10 exercise hie Hingis Treuly’ (Enger, 31). Parale, on the other lism, i he “execiniv? see af language’ and Jor Saussure involves both “the eambination by whith the speaker usrs the code of the linguistic systems in onder 10 express his oss thougbs* and she peyeno physical mechanisms whiele permit hin) to °9