Klasterinė tipologinių kalbos ritmo grupių analizė

168 ac a Linguis ica Li huanica LXIX agnė PukeViČiū ė Vy au o didžiojo uni e si e as The di ec ions o scien i ic esea ch: he peculia i ies o he Li huanian language. as a kaZLauskienė Vy au o didžiojo uni e si e as Resea ch a eas: esea ch on he Li huanian language: acous ic sound p ope ies, phono ac ics, whis ling sounds, no ms and cus oms, hy hm o speech; he applicable dialec . This a icle seeks o answe wo ques ions: 1) is clus e analysis good a inding hy hms o ypological simila i ies and on he basis o i i is possible o classi y languages no analyzed un il now and 2) o which clus e […] he skiemen's o qua e hy hm o languages […] belongs Li huanian and clus e analysis o ypological g oups o Language hy hm La ian. The s udy in ol ed i e na i e speake s (La ian, Li huanian, Swedish, Icelandic and Chinese) ansla ing and eco ding he same ex o abou 1 minu e. The s udy was unded by he Li huanian Science Council (ag eemen no. Li -5-4). His ini ial esul s a e summa ized o hose who helped o segmen he esea ch ma e ial: p o . habil. d . Pe e Vanagui, Valge ðu bja nadó i , Magnus Liw, huashan Li and doc. D . aauš ai hanked me o helping o p epa e he ma e ial o he hy hm co ela es, he clus e analysis, he Li huanian, La ian, Swedish, Icelandic, Chinese languages. agnės Puke ičiū ės magis o baigiamajame da be (Akus iniai lie u ių i la ių kalbų i mo ko elia ai, 2013, ado ė p o . d . as a kazlauskienė). s aipsnio au o ės nuoši džiai dėkoja in o man ams, į ašiu- s a is inei analizei i pa a imus di ban su sPss p og ama. esMiniai ŽodŽiai: i mas, ki inio i mo kalbos, skiemeninio i mo kalbos, akus iniai kal- keyWo ds: hy hm, s ess- imed languages, syllable- imed languages, acous ic co ela es o hy hm, clus e analysis, Li huanian, La ian, Swedish, Icelandic, Chinese. klas e inė ipologinių kalbos i mo g upių analizė 169s aipsniai / a icles The p og am Co ela o e (2.1 e sion) was c ea ed o calcula e acous ic hy hm co ela es. Clus e analysis is a me hod used by esea che s o de e mine he a e a which a language speaks. he da a ha e been compiled in o an sPss p og amme and clus e analysed. does no a lik o y imo ezul a ai odo, kad hie a chinė i d iejų žingsnių klas e inė analizė iena- p o ide meaning ul answe s. Hie a chical analysis is qui e good o clus e ing less s a is ically p ocessed a iables (ΔV, Δc, %V, pa ly also a coΔV, a coΔc). The wo-s ep analysis should be used wi h g ea cau ion when classi ying languages by hy hm, because i s esul s a e con adic o y. This a icle aims o answe wo ques ions: 1) whe he he clus e analysis is handy o inding ka u su daugeliu ki inio i mo kalbų. odėl galima bū ų da y i a sa gią iš adą, kad abiejų gy ųjų bal ų kalbų i mas panašus i jį lemia ki is. ai ki inio i mo kalbos. ypological hy hm simila i ies and on he basis o i so a unanalysed languages can be classi ied 2) which clus e […] s ess- imed o syllable- imed languages […] Li huanian and La ian languages all in o. Fo his s udy, i e na i e speake s (La ian, Li huanian, Swedish, Icelandic and Chinese) ansla ed and eco ded he same ex . ex du a ion was abou 1 minu e. eco ds we e anno a ed wi h he p og am P aa and acous ic co ela es o hy hm we e calcula ed using he p og am Co ela o e de eloped by Paolo Mai ano (2.1 e sion). esul s o o he languages es ima ed by he au ho o he p og am Mai ano (2011) was used o clus e analysis. he da a was en e ed in o he sPss p og am and a clus e analysis was pe o med. The esul s show ha he hie a chical and wo-s ep clus e analysis does no p o ide unambiguous esul s. hie a chical analysis is p e y handy o clus e less s a is ically p ocessed a iables (ΔV, Δc, %V, pa ly a coΔV, a coΔc). wo-s ep analysis should be used wi h cau ion because o i s con o e sial esul s. Li huanian and La ian, om all he a iables, all in o he same clus e , along wi h a numbe o s ess- imed languages. The e o e, i can be ca e ully concluded ha he hy hm o he wo Bal ic languages is simila and caused by s ess. They' e s ess- imed languages. Lee Pike (1945) is one o he i s o ex ensi ely discuss he hy hmic hy hm o syllable- imed languages, based on many la e 20 h-cen u y s udies based on Pike's hypo hesis o he ypological classi ica ion o language hy hm. By measu ing and i miškumo ipologinius ski umus. jis sky ė: 1) ki i nio i mo (ang. s ess- imed) kalbas, ku ių i mo pag indas y a ki is (pa yzdžiui, anglų, okiečių kalbos), compa ing he du a ion o acous ic skies and in e mi en in e als, he aim was 170 ac a Linguis ica Li huanica LXIX These wo ypological g oups we e la e agnė PukeViČiū ė, as a kaZLauskienė joined by a hi d […] mo i nio hy hm g oup, which o ms he basis o he panašios kiekybės skiemenų a osena (pa yzdžiui, p ancūzų, ispanų kalbos). P ie hy hm o he languages o which i belongs: mo a (languages, Japanese, Tamil). o jus i y he exis ence o wo hy hm g oups and hei di e ences (abe c ombie 1967; oach 1982). Un o una ely, expe imen al s udies ha e no con i med he isoch onism o he sound segmen s - he epe i ion a equal in e als o he app oxima ion o hei quan i ies. So we ied o look o indica o s o he hy hm o o he languages. Rebecca McKee e Daue (1983), based on he ac ha he phonological sys em o language di ec ly de e mines he hy hm, p oposed o di ide languages in o hy hm g oups acco ding o he p esence o absence o ce ain phonological cha ac e is ics. I 's he educ ion o o es and he complexi y o he skele al s uc u e. Based on his idea and applying s a is ical me hods, F anckus Ramus, Ma ina Nespo and Jacques Mehle is (1999) ied o de e mine he acous ic hy hm co ela es o speech ( a iables ob ained by calcula ing he leng h o owel and consonan in e als): ΔV […] he s anda d leng h o owel in e als, Δc […] he s anda d leng h o consonan in e als, %V […] he leng h o owel in e als mus be highe han ΔV and Δc, and %V […] lowe han he hy hm o consonan p opo ciją. Minė i au o iai hipo e iškai spėjo, kad ki inio i mo kalboms būdingas didesnis balsių ukmės a ian iškumas i sudė ingesnė skiemens s uk ū a, odėl in e als. Un o una ely, he esul s o he speci ic languages hey analyzed only pa ially con i med his. Fu he s udies ha e poin ed o he possible in luence o a ious linguis ic and non-linguis ic ac o s on hy hm. In o de o no malize he empo di e ences, Es he G ab and Ling ee Low (2002) calcula ed he pai a iance indices: nPVi […] ocal in e als, PVi […] ocal in e al index. I is expec ed ha he a iabili y o he leng h o owels and consonan s is high o he syllabic hy hm languages, hei nPVi and PVi a e high, and hese indices o he syllabic hy hm languages a e low. Volke is dellwo and Pe a Wagne (2003) also sough o a oid he in luence o he empo in e up ion. hey mul iplied he s anda d de ia ion o he leng h o owel and consonan in e als by 100 and calcula ed he coe icien s o a ia ion: a coΔV (consonan in e als) and a coΔc (consonan in e als). Howe e , nei he he pai ed a iance indices no he a ia ion coe icien s help o de ine he hy hm o ypological simila i ies and o di ide languages in o ypological classes. Howe e , he me hodology o se ing he hy hm pa ame e s discussed has been popula o qui e some ime, and is s ill popula oday, al hough i has ecei ed jus i ied c i icism (in he case o A ani i 2012)2. 2 The e has al eady been a b oade discussion o he me hods o s udying he hy hm o language and he i mą ž . kazlauskienė 2013. klas e inė ipologinių kalbos i mo g upių analizė 171 a icles by a icles The second me hod is he me hod o clus e analysis, which assesses he co ela ion Šiam y imui pasi ink a išbandy i naują būdą – ekspe imen iškai pa ik in i akus- be ween he hy hm o he wo ds and he simila i y o he language in gene al. In s a is ics, i is used o de e mine he simila i y o objec s o phenomena and o classi y hem in o clus e s (i.e. g oups o simila objec s). The aim o clus e analysis is o g oup he a iables unde s udy in such a way ha di e ences wi hin he clus e a e as small as possible and be ween clus e s a e as la ge as possible (Vai ke ičius, sauda gienė 2006). I'm hoping ha by doing a clus e analysis, we'll be able o cla i y wo hings ha a e he pu pose o his s udy. 1. do he acous ic hy hm co ela es o he Swedish, Icelandic and Chinese languages calcula ed by Paolo Mai ano (2011) and o his s udy all in o he same clus e s? Do he co ela ions be ween Mai ano and ou calcula ion o he hy hm o Swedish and Icelandic di e om he hy hm o Chinese? I hese ques ions a e answe ed in he a i ma i e, hen clus e analysis is eally good a inding hy hms o ypological simila i ies and i can be used o classi y languages ha ha e no ye been analyzed. 2. he second conce n is, which clus e o […] linguis ic o hy hmic languages […] does he Li huanian and La ian languages belong o? The acous ic hy hm co ela es o Swedish, Icelandic and Chinese languages ha e al eady been calcula ed by o he au ho s ( his s udy is based on da a om a publicly a ailable sou ce on he In e ne […] Mai ano 2011). The i s wo languages a e assigned o he ki an hy hm, he Chinese […] o he skiemen hy hm g oup. The hy hm o Li huanian and La ian languages acco ding o his me hodology has no ye been s udied. In ea lie s udies o he hy hm o he Li huanian language, he basis o hy hm was conside ed o be he ki is, bu no he du a ion o he acous ic uni s, bu he dis ibu ion o he ki is in a ious segmen s. The hy hm o he eiliou ic language was s udied in his way by Juozas Gi dzijauskas (1986, 1979) and Leona das Sauka (1974, 1978), and he hy hm o he p ose and p o e bial language by Vale ia Ramonai ė (1960, 2006), boni acas s undžia (1991) and as a kazlauskienė, inga okai ė (2005). possibili y o applying hem in desc ibing he hin s o he Li huanian language abou he hy hmic na u e o he La ian language. Zinny sans bond, dace Ma kus and Ve na s ockmal (2003) compa ed na i e La ian and na i e Russian speake s, speaking La ian, wi h a p ose: con as ing speech du a ion, hy hm and hy hm. hei a icle does no p o ide a de ailed analysis and in e p e a ion o he hy hm o he La ian language, al hough Ramuso, Nespo , Mehle (1999) and G ab and Low (2002) sugges sizes. In he s udy o bond, Ma kus and S ockmal (2003) i was assumed ha La ian languages a e cha ac e ized by a skiemenic hy hm. The s anda d o he In e na ional Phone ic Associa ion (SUN, No h Wind and Sun) is al -usage-english.o g/no h_ x .h ml) siūlomas eks as (ang. The No h Wind and he used no only by hy hmologis s bu also by all phone icians. when analysing ma e ial o a di e en na u e, i would be inapp op ia e o compa e he esul s o his s udy wi h da a om o he languages ob ained on he basis o he abo emen ioned ex . he ac a agnė PukeViČiū ė, as a kaZLauskienė Linguis ica Li huanica LXIX 172 Fi e na i e languages (La ian, Li huanian, Swedish, Icelandic and Chinese) ansla ed and ead he ex . The in o man 's speech was eco ded on a ape eco de in a quie oom. Each eco ding is sligh ly sho e han one minu e. The s udy ma e ial was p ocessed and he empi ical da a ex ac ed in se e al s ages: 1. consul ing wi h he in o man o each language, he P aa p og am was anno a ed eco ds. The anno a ion is based on bo h he image o he oscillog am and he hea ing. he ex anno a ion g id o he p og am (Tex G ide) is sa ed in * x o ma . Using he Co ela o e (2.1 e sion) p og am de eloped by Mai ano (2011) a he A u o Gene e Expe imen al Phone ics Labo a o y in Tu in, I aly, we calcula ed he s a is ical co ela ions o he acous ic hy hms o ou eco dings ( esul s show ha he esul s kiami 1 len .). klas e inei analizei, kaip jau minė a, panaudo i duomenys iš Mai- ano (2011). jie eikalingi am, kad bū ų galima inkamai a lik i klas e inę analizę, a e much mo e eliable when a la ge amoun o da a is g ouped. This is Len e. Acous ic hy hm co ela es and he impo ance o s a is ical quan i ies4 kalba Δc %V ΔV a coΔV a coΔc nPVi PVi Li huanian 52 54 54 54 66 57 57 In he case o he Uni ed Kingdom, he Uni ed Kingdom, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland, I eland and I eland, he applican mus ha e been esiden in he Uni ed Kingdom o a leas one yea . o Po ugal 54 49 55 48 44 57 56 F ench 40 49 40 45 41 45 44 Finnish 47 48 45 47 46 45 55 G eek 37 47 34 50 48 52 42 English, F ench, Ge man, F ench, Ge man, F ench, Ge man, F ench, F ench, Ge man, F ench, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Ge man, Es onian 3 All in o man s […] people wi h highe philological educa ion, wo king o con inuing hei s udies a uni e si ies: La ian and Icelandic […] lec u e s, Swedish […] head o s a , Li huanian and Chinese […] pos g adua e s uden s. He e and he e, he eco ds o ou s udy a e ma ked wi h an index o 2, e.g. Swedes 2, and Mai ano's da a […] wi hou an index. klas e inė ipologinių kalbos i mo g upių analizė 173s aipsniai / a icles kalba Δc %V ΔV a coΔV a coΔc nPVi PVi A abic 55 45 47 56 55 47 64 i alų 54 44 44 53 49 48 62 olandų 46 43 45 62 50 68 53 Danish 61 42 40 46 52 50 66 Spanish 52 42 30 42 53 42 60 Russian 55 41 34 43 48 45 63 English 59 40 48 62 51 61 70 Polish 66 36 32 41 49 41 72 3. agg ega e he da a in o he sPss p og amme (see mo e a h p: www.insol.l ) o s a is ics/, Vai ke ičius, sauda gienė 2006) i a lik a jų klas e inė analizė. nag inėjan lie u ių i la ių kalbas sPss p og ama, bu o a lik a hie a chinė i a wo-s ep clus e analysis. In a hie a chical analysis, he numbe o clus e s is no known in ad ance. clus e s a e o med on a hie a chical basis […] all da a is kep in one clus e composed o smalle g oups. hey may be es ablished by di iding one common clus e in o smalle ones o by joining smalle clus e s in o la ge ones. The wo-s ep analysis di e s om he one discussed in ha he numbe o clus e s is known in ad ance (Vai ke ičius, sauda gienė 2006). o choose wo clus e s in he wo k, because wo language hy hm g oups […] ki nian and ečioji, mo inio i mo, g upė neį auk a į analizę, nes nekel a hipo ezė, kad bal- ų kalbos galė ų p iklausy i šiai g upei. a liekan klas e inę analizę, objek ai pe - skiemennian a e examined. The simila i ies wi hin a clus e become much g ea e han he simila i ies be ween clus e s (Vai ke ičius, sauda gienė 2006). The numbe zul a ai ik inami aikan ik d iejų žingsnių klas e inę analizę, nes pasi ink as mažas nag inėjamų objek ų skaičius (šešios kalbos). be o, iš anks o žinomas klas- o Swedish, Icelandic and Chinese e ie e s and he hy hm g oup o which hese languages a e assigned ( wo hy hm g oups and he esul s o wo su eys). In o de o ind ou whe he Mai ano and he Swedish, Icelandic and Chinese languages eco ded in his s udy a e simila in size and all in o he same g oup, a clus e was ca ied ou (25,55,7,8,8,8,46,48,83,210,7,720) om he able ha he mean o s anda inis p iebalsių ukmės nuok ypis (Δc) i balsinių in e alų ukmės p ocen as (%V) he i s clus e is highe han he a e age o he second clus e , and he clus e is highe han he a e age o he second clus e , bu in his case he e is a ai ly la ge 174 ac a Linguis ica Li huanica LXIX clus e analysis in wo s eps. Fi s , he mean alues o agnė PukeViČiū ė, as a kaZLauskienė he quan i a i e a iables o each clus e and hei s anda d de ia ion a e calcula ed (see Table 2). 2 The Len e. clus e cen es by pe cen age o o ing in e als and s anda d de ia ion o o ing in e als % o Vc Vidu kis s anda inis nuok ypis Vidu kis s anda inis nuok ypis de ia ion. This means ha he i s clus e is cha ac e ized by a mo e complex 145,00 3,916 58,00 5,715 skiemens s uc u e (i.e. mo e p obalsi skiemenyje (Δc highe )), which also de e mines a lowe pe cen age o ocal in e als (%V). The esul s o his analysis indica e ha he mos simila %V and Δc ha e one clus e o he islandic (%V = 40, Δc = 60), islandic (%V = 47, Δc = 52), Swedish (%V = 44, Δc = 55), Swedish (%V = 49, Δc = 65) languages belonging o he hy hm g oup, and he o he clus e o he Chinese (%V = 51, Δc = 44) and Chinese (%V = 62, Δc = 48) languages. Thus, %V and Δc a e ai ly eliable indica o s by which he languages analysed ac ually di e and which do no depend on he ac ions o esea che s o in o man s' objec i es (speaking pace, in ona ion, e c.) o subjec i e ones (w i ing quali y, anno a ion di e ences, e c.). The coe icien s o a ia ion a e qui e signi ican ly di e en be ween he wo clus e cen es acco ding o he coe icien s o a ia ion. Table 3 shows ha he a e ages o a coΔV and a coΔc o he second clus e a e highe . 3 The Len e. clus e cen es by a ia ion coe icien s a coΔV a coΔc Vidu kis s anda inis nuok ypis Vidu kis s anda inis nuok ypis he clus e 2 142,67 3,215 50,33 4,933 50,33 0,517 62,67 3,055 bend as 46,50 4,680 56,50 7,688 klas e inė ipologinių kalbos i mo g upių analizė The esul s o his s udy a e e y unexpec ed: he mos simila a ian s o a coΔV and a coΔc a e ound in he i s clus e o Mai ano (2011) analysed islande s ( a coΔV = 45, a coΔc = 56), Swedes ( a coΔV = 39, a coΔc = 48) and Chinese ( a coΔV = 44, a coΔc = 47) languages. All he eco ds we s udied all in o he second clus e : Icelandic 2 ( a coΔV = 50, a coΔc = 60), Swedish 2 ( a coΔV = 50, a coΔc = 66) and Chinese 2 ( a coΔV = 51, a coΔc = 62). Thus, he a ia ion coe icien , which should be mo e objec i e in he ypological classi ica ion o languages, c ea es mo e con usion and sugges s ha i s esul s depend hea ily on he expe ience o anno a ion in a pa icula language. The a ia ion coe icien s o ou eco ds a e simila , which sugges s ha pe haps i is no enough o consul na i e speake s when anno a ing, bu he anno a ion i sel needs o be done by na i e phone icians. The di e ence in he pai a iance index be ween he wo clus e cen e s acco ding o nPVi and PVi is qui e conside able (see. 4 o his Regula ion. ). 4 The Len e. clus e cen es by pai a iabili y indices nPVi PVi Vidu kis s anda inis nuok ypis Vidu kis s anda inis nuok ypis he 148,00 3,000 58,80 6,458 clus e 2 60,00. 82,00 . The o al is 50,586 (6,667), bu language g ouping is e en mo e ambiguous han he a ia ion coe icien s. The i s clus e consis s o all languages excep Swedish 2, which was analyzed o his s udy (i s PVi is e y high […] 82, al hough nPVi is simila o o he languages […] 60) […] i is he only one in he second clus e . The pai a ia ion indices o o he languages a e e y simila : Icelandic (nPVi = 45, PVi = 63), Icelandic 2 (nPVi = 51, PVi = 56), Chinese (nPVi = 48, PVi = 54), Chinese 2 (nPVi = 51 PVi = 53), Swedish (nPVi = 45, PVi = 68). In his case, i is clea ha wo-s ep clus e analysis is no use ul, because he i s clus e needs o be b oken down in o smalle g oups. Summing up he esul s o he wo-s ep clus e analysis o h ee languages, i can be said ha he s anda d de ia ion o owel leng h (Δc) and he pe cen age o owel in e al leng h (%V) is a ai ly eliable indica o o g ouping languages in o classes. Two-s ep clus e analysis o a ia ion coe icien s and pai a iance indices in his case aises mo e ques ions han answe s. 176 ac a agnė PukeViČiū ė, as a kaZLauskienė Linguis ica Li huanica LXIX 4. Lie uVių i La Vių kaLbų i Mas In o de o ind ou which ypological hy hm g oup belongs o he La ian and Li huanian languages, a hie a chical clus e analysis was i s ca ied ou o check how he i balsinių in e alų ukmės p ocen as (%V) languages a e g ouped acco ding o ΔV, %V and Δc and o which clus e he Li huanian and La ian languages belong. The esul s o he hie a chical clus e analysis ob ained om he da a o he o he languages used o his analysis show ha he Li huanian, La ian, Ge man, A abic, English, Czech, I alian and Po uguese languages ha e he mos simila (2011: 170, 179, 199). alues o V, V and C, hus o ming a single clus e , wi h he V anging om 44 o 58, he V om 40 o 54, and he C om 52 o 66. In his clus e , only he I alian language is conside ed o be closes o he skiemen hy hm (as can be seen, he clus e analysis seems o doub his), all o he languages (excep , o cou se, he analyzed Bal ic languages, whose hy hm ype is unknown) belong o he hy hm o he qua e . The second g oup was made up o many o he languages wi h he same hy hm, bu i included a numbe o hose ha ended o be conside ed o be o a mixed o mixed hy hm (see box 5). The ΔV o hese languages anges om 30 o 45, he %V […] om 36 o 51, he Δc […] om 37 o 66. Looking a he esul s, one hing is ce ain: he i s clus e did no include a single p o o ype o a skiing hy hm language (F ench, Spanish, G eek), and he second clus e had only wo p o o ypes o a skiing hy hm language (Russian, Du ch). As can be seen om able 5, Li huanian and La ian all in o he i s clus e and a e mo e simila o he p o o ypical hy hmic languages ( he acous ic co ela es o he languages a e gi en in able 1). The 5 h Len e. The esul s o he hie a chical clus e analysis a e based on he s anda d nuances and p opo ions o he owels, he nume ical hy hm o he clus e , he hy hm o he g oup and he language, he hy hm o he chan , he hy hm o he ski , he hy hm o he ski . 1English, Ge man, A abic, Po uguese, Czech, I alian, Li huanian, La ian 2Russian, Du ch, Icelandic, Tu kish, Danish, Polish, Swedish, Chinese, Es onian, Finnish, Spanish, F ench, G eek, 5 P o o ypical speci ic hy hm g oups o languages men ioned in his and o he ables, he hy hm hy hms). o o he languages is discussed (some o which a e e en conside ed ei he non-classical o mixed klas e inė ipologinių kalbos i mo g upių analizė 183 he a icles o he no he n wind blow as as as he wind, and he e o e as e han he wind blow, so ha he hickness o he a ele 's wais is co e ed wi h a seal, and a he end he no he n wind blows up o y. Þá ó sólin að skína, og undi eins ó e ðamaðu inn ú akkanum. Þá a ð I'd like o acknowledge ha he sun is s onge han he moon. 6. he Chinese ex s in he language you yi ian, dang yi ge guo zhe da yi de l xing zhe chu xian de shi hou, bei eng he ai yang zheng jiu zheng chao zai jing shui geng jiang. This is he i s ime ha he Chinese language has been ansla ed in o Chinese. I don' know wha you' e alking abou . I don' know wha you' e alking abou . I'm going o ake a look a i . I'm going o ake a look a i . I'm going o ake a look a i . Zui Hou, I'm going o ake a look a Feng Feng Qi Le. They' e like, "Oh, my God, hey' e like, "Oh, my God, hey' e like, "Oh, my God, hey' e like, "Oh, my God, hey' e like, You yi, and eng bu de bu cheng en ai yang bi a geng qiang. This is he i s ime I' e seen his. suMMa y his clus e analysis o ypological g oups o Language hy hm a icle aims o answe wo ques ions: 1) whe he he clus e analysis handy inding languages 2) in which a clus e s ess- imed o syllable- imed languages Li huanian and hy hm ypological simila i ies, and i can be classi ied on he basis o i so a unanalysed La ian languages all in o. The s anda d ex , p oposed by he In e na ional Phone ics Associa ion (see h p: www.usageenglish.o g) (The No h Wind and he Sun, Li h. Fi e na i e speake s (La ian, Li huanian, Swedish, Icelandic and Chinese) ansla ed and eco ded i (each eco d was much sho e han 1 min). The eco ds we e anno a ed by he p og am P aa and acous ic co ela es o hy hm o eco ds ha we ha e done ha e been calcula ed using he p og am Co ela o (2.1 e sion) which was de eloped by Pado Mai ano (2011) in A u o Gene e Expe imen al Phone ics Labo a o y o Tu in (I aly). I was ound ha Li huanian Δc=52, %V=54, ΔV=54, a coΔV=54, a coΔc=66, nPVi=57, PVi=57; La ian Δc=55, %V=52, ΔV=58, a coΔV=55, a coΔV=55, a coΔV=51, a coΔV=61, a coΔV=61, a coΔV=61, a coΔV=61, a coΔV=51, a coΔV=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=51, a coVi=62, Vi=63, Vi=62, Vi=6 and so on and so on and so on and so on and so on. 184 agnė PukeViČiū ė, as a kaZLauskienė ac a Linguis ica Li huanica LXIX da a o o he languages es ima ed by he au ho o he p og am Mai ano (2011) was used o clus e analysis The da a was en e ed in o he sPss p og am and a clus e analysis was pe o med. he esul s show ha he clus e analysis (bo h hie a chical and wo-s eps) does no p o ide unambiguous esul s. Hie a chical analysis is qui e handy o clus e less s a is ically p ocessed a iables (DV, Dc, %V, pa ly a coDV, a coDc), a wo-s ep analysis migh be possible o ely on, and wi h mo e complex indices (PVi). Howe e , a wo-s ep analysis is begging he ques ion, because i does no lea e he hi d op ion (s ess- imed o syllable- imed, canno be mixed). Li huanian and La ian, acco ding o all he a iables, all in o he same clus e oge he wi h a numbe o s ess- imed languages. The e o e, i can be ca e ully concluded ha hy hm o he wo bal ic languages is simila and caused by s ess. hey a e s ess- imed languages. Clus e analysis (in pa icula , Swedish, Icelandic, Chinese language compa ison be ween he wo ypes o da a) has shown ha he anno a ion p ecision is pa icula ly impo an in he esea ch o hy hm. on he o he hand, he indings a e ele an no o he ypological classi ica ion o languages' hy hm and no o he indica o s o languages' hy hm. They can be as guidelines o u he esea ch o ano he excuse o ques ion he exis ence o s ic ly de ined classes o speech hy hm. Deli e ed on 11 July 2013 agnė PukeViČiū ė Vy au o Didžiojo uni e si e as This Regula ion shall en e in o o ce on he day ollowing ha o i s publica ion in he O icial Jou nal o he Eu opean Union. [Email p o ec ed] as a kaZLauskienė Vy au o Didžiojo uni e si e as This Regula ion shall en e in o o ce on he day ollowing ha o i s publica ion in he O icial Jou nal o he Eu opean Union. [Email p o ec ed]