.\" groff -tme .nh .ds T "A Tagset for Manx .de $0 .if '\E$3'1' .sp .5 .. .de $1 .sp .5 .. .de fe\" display a feature name \f[HB]\E$1\fP\E$2 .. .\"ig .nr pp 11 .nr sp 13 .nr qp 10 .\". .ig .nr pp 14 .nr sp 16 .nr qp 13 .nr fp 11 .. .fp 5 CR\" Typewriter on font position 5 .fp 6 HR\" Sans serif on font position 6 .(l C .ps +8 .b "\u\u\*T .ps 12 John D. Phillips .ps 10 .)l .\"he 'John D. Phillips'\*T, \fIdraft of \n(dy/\n(mo/\n(yr\fP'%' .\"he 'John D. Phillips'\*T'%' .eh '%'\*T'John D. Phillips' .oh 'John D. Phillips'\*T'%' .de TS .sp .5 .ps +1 .vs +4 .. .de TE .sz \n(pp .sp 1 .. .lp The language to which this tagset relates is the spoken Manx of twentieth century native speakers. The tags are derived from an analysis of a corpus of transcribed speech. The Manx of the Bible and other 18th century texts which make up most of the available written Manx, is very different to that of the spoken language of the 20th century due to the natural processes of linguistic change over the centuries (also partly to the difference of register). However, the tagset presented below could be adapted fairly easily for use with the older texts, and suggestions are made to that effect in the comments on the various parts of speech below. .pp A tag is made up of a capital letter expressing the part of speech followed by letters and numbers showing the values for the various .i features associated with that part of speech. For each part of speech below, a table shows the range of values for each feature. Some features or values presuppose other features or values, and these dependencies are shown by horizontal lines dividing the tables. The symbol \(es shows where a feature is not relevant. The notation .i "{a,b} means .i a or .i b , so that N{b,l}{c,v} means Nbc, Nlc, Nbv, or Nlv. .sh 1 "Noun .TS box,center; l l l. POS mutation type _ N b=base c=common l=lenition p=proper v=verbal .TE No attempt is made to distinguish noun-like and verb-like uses of verbal nouns (cf. Leech & Wilson's English tagset, which does not distinguish participles and gerunds). .pp Some nouns (including some verbal nouns) have a plural form so that singular and plural could be distinguished, but there is no syntactic number agreement, and since anaphoric number agreement is semantic, a morphological number marker would be positively misleading. .pp Older Manx had syntactic agreement, and anaphoric agreement was probably morphological: number {singular,plural}, gender {masculine,feminine} and case {nominative,genitive,dative} would need to be distinguished. A few genitive forms have actually survived but are here (following Cregeen, 1835) classed as adjectives, tag A{b,l}a. .bp .sh 1 "Verb .TS box,center; l l l l. POS dependency type person-number _ V i=independent a=auxiliary 0=none d=dependent b=main \fIbe\fP M=1sing m=modal S=1plural _ _ p=personal {m,o,y,i,s,u,a,0} s=personal \fIs-\fP (see pronoun) _ _ S=impersonal \fIs-\fP \(es d=defective past o=other finite .TE The part of speech `verb' is a closed class comprising auxiliaries, modals and the like. The productive class of verbal nouns is included under nouns, tag N{b,l}v. Since choice of tense seems to be free in all syntactic environments, and since different types of verb distinguish a different range of tenses, tense is not marked on the tags. .pp Mutation is not relevant to verbs: though many of the forms classified here are historically the results of mutation, this has no synchronic relevance and is not productive or paradigmatic in the modern language. .pp .fe Independent forms are those used to begin a sentence, e.g. .i "ta, yinnagh, foddagh, seinn" ; .fe dependent forms follow .i cha and other preverbs, e.g. .i "vel, jinnagh, noddagh, nheinn" . .pp The values for .b type are: .in 1m .ti 0 .fe Auxiliaries (V{i,d}a0 forms): .i "nee, jean, bee, ta, te, vel, va, ve, row, ren, yinnagh, jinnagh, veagh, beagh" . .ti 0 Forms of the verb .fe be used as a main verb: .i "bee, ta, te, vel, va, ve, row, veagh, beagh" . .ti 0 .fe Modals (Vim0 forms): .i "foddagh, lhiddagh" . .ti 0 Verbs unspecified for tense but with .fe personal , inflections e.g. (Vipm forms): .i "er-lhiam, [cha] saym" . .ti 0 Verbs in .fe s- inflecting for person and two tenses, e.g. .i "saillym, baillym, sailt, bailt. .ti 0 Verbs in .fe s- inflecting for two tenses only, e.g. (ViS forms) .i "seinn, beinn, share, bare, slhiass, blhiass". .ti 0 .fe "Defective past tense forms (Vid): .i "dooyrt, hooar, hie" . .ti 0 Tag V{i,d}o is provided to cover cases where otherwise obsolete forms turn up in set phrases and the like. .pp There are two sets of values for .b person-number .in 1m .ti 0 {S,M}: auxiliaries, .i be , and modals occasionally have a clitic first person pronoun attached, e.g. .i "tamain, neemayd, yinnin, lhiddym, foddym" . .ti 0 {m,o,y,i,s,u,a}: personal verbs inflect for person-number with the same endings as prepositions. .pp In older Manx, any verb could have inflected finite forms. It should be adequate to tag these as V{i,d}o{0,S,M}. .sh 1 "Adjective .TS box,center; l l l l. POS mutation type degree _ A b=base o=ordinary p=plain l=lenition c=comparative e=equative _ _ v=verbal \(es d=demonstrative a=adj. derivative .TE Regular comparative/superlative forms (Aboc) are made by prefixing .i s or .i s' to the plain form, e.g. .i "s'aalin, s'bwaagh", and there are a few irregular forms e.g. .i "share, smoo, sudjey" . A few adjectives have special equative forms (Aboe): .i "hammah, choud, wheesh" . Plain adjectives in equative constructions should be tagged Abop. .pp .fe "Verbal adjectives (A{b,l}v) have the suffix .i -t or .i -it , e.g. .i "curt, jeant, jaroodit" . .fe "Demonstrative adjectives (Abd) are: .i "shoh, shen, shid" . .fe "Adjective derivative (A{b,l}a) is Cregeen's (1835) name for what is historically the genitive case of a noun. This is a tiny, unproductive class of adjectives, whose members could probably be subsumed under A{b,l}op (they are not gradable, but then neither are some ordinary adjectives, e.g. .i "un, ynrican" ). An example is .i geayee `wind' in .i "muillin geayee" `windmill'. .pp Older Manx had plural adjectives, agreeing with plural nouns, and the adjective derivative was a productive part of the nominal case system. .bp .sh 1 "Pronoun .TS box,center; l l l l l l. POS mutation type function person-number place _ P b=base p=personal n=nominal m=1s (\fImee\fP) w=word l=lenition o=2s (\fIoo\fP) e=enclitic y=1sn (\fIyh\fP) e=3sm (\fIeh\fP) i=3sf (\fIee\fP) s=1p (\fIshin\fP) u=2p (\fIshiu\fP) a=3p (\fIad\fP) _ _ p=possessive {m,o} \(es _ c=complement \(es _ d=demonstrative \(es i=interrogative t=indefinite x=emphatic/reflexive r=reciprocal 2=dual .TE The feature mutation may be unnecessary. The only pronouns which it would be syntactically and phonetically possible to lenite are the indefinite and reciprocal pronouns. In the corpus on which this tagset is based, there are no examples of lenition of indefinite pronouns, and plain and lenited forms of the reciprocal pronoun seem to be in free variation. .pp The plain .fe personal pronouns Pbpn{m,o,y,e,i,s,u,a}w are .i "mee, oo, yh, eh, ee, shin, shiu, ad" . In the third-person singular, the pronoun spelt .i a by Kneen, .i yh in some older texts, is used but the forms .i "eh, ee can optionally be used to refer to (biological) males and females respectively. .fe Enclitic pronouns .i "s, s, syn, sh, yn, ish, syn can be suffixed to lexical pronouns, nouns, and inflected verbs for emphasis or contrast. For consistency, the `emphatic pronouns' .i "mish, uss" , etc. are treated as plain pronoun plus enclitic despite their irregular spellings. .pp The .fe possessive pronouns Pbpp{m,o} are .i my and .i dy/dty (first and second person singular) only. The predicative pronoun Pbpc, introducing a .fe complement is .i ny only, e.g. in .i "\(lqtraa va mish .bi ny .i lhiannoo.\(rq .pp Other values of the type feature are: .in 1m .ti 0 .fe demonstrative pronouns Pbd, .i "shoh, shen, shid" ; .ti 0 .fe interrogative pronouns Pbi, .i "quoi, cre, c'" ; .ti 0 .fe "indefinite pronouns Pbt, .i "fer, follee, sthoo" (singular, plural, mass) used for type anaphora; .ti 0 .fe emphatic/reflexive pronoun Pbx, .i "hene" (only); .ti 0 .fe reciprocal pronouns P{b,l}x, .i "killey, chilley \(em is the lenition syntactically important? .ti 0 The .fe dual form .i ny-niss is given a separate tag Pb2 because it is not clear where else it could go. .pp In older Manx, the possessive and predicative pronouns made more distinctions than in the modern language: there is a third-person singular undifferentiated for gender and a plural undifferentiated for person (\fInyn\fP), so that the range is Pbp{p,c}{m,o,y,n}. For many texts, Pbpny will not be needed. .sh 1 "Article .TS box,center; l l. POS number _ T g=general p=plural .TE The definite article Tg is .i "yn, 'n, y" , the choice determined by the phonetic environment. A special plural article Tp .i ny turns up in a few set phrases, e.g. .i "beeal ny baareyn `a crossroads'. There is no indefinite article. .sh 1 "Adverb .TS box,center; l l. POS modifies _ R c=clause a=adjective v=verb t=topic .TE Adverbs are a heterogenous class, here subdivided according to what they modify. .in .ti 0 .fe Clause -modifying adverbs Rc refer to time, such as .i "hoshagh, jea, jedoonee" , or to place, such as .i "cooidjagh, ersooyl, magh" . .ti 0 .fe Adjective -modifying adverbs Ra include intensifiers such as .i "ro, fee(r), brouw, foddey, ram" , also equative .i cha , and .i dy marking adverbial use of an otherwise unmodified adjective. These intensifiers precede the adjective they modify. Ordinary adjectives used to intensify a preceding adjective should get the tag Abop, e.g. .i agglagh in .i "ching agglagh" . .ti 0 .fe Verb -modifying adverbs Rv comprise the pre-verbal particles negative .i cha , and direct interrogative .i "nagh, ny" ; other pre-verbal particles (including subordinating .i "nagh, ny" ) are classed as subordinating conjunctions Cs in conformity with the tagging guidelines. .pp The .fe topic markers are .i "ne interrogatively and after .i cha ; .i re elsewhere, e.g. .i "Insh daue dy .bi re .i "mish ren cur yh daue" . The historically related .i she is classed as an interjection, tag I. .pp Subdivision of Rc based on syntax is possible. .bp .sh 1 "Preposition .TS box,center; l l l. POS inflexion person-number _ S i=declinable {m,o,y,i,s,u,a,0} a=auxiliary (see pronoun) _ 0=indeclinable \(es .TE .fe Declinable prepositions have a plain form Si0 and inflected forms for each person-number combination, except that there are no specific masculine forms (i.e. the tag Sie does not exist): examples are .i "lesh, da, rish, marish" , all four of which can be either Si0 or Siy. Some .fe indeclinable prepositions S0 inflect for person-number with the aid of an .fe auxiliary , e.g. .i "son-aym, eddyr-ock, mygeayt-y-mysh where the inflected auxiliary prepositions .i "ec, mysh supply person-number for the indeclinable prepositions .i "son, eddyr, mygeayt" . The sequence .i "son-aym therefore gets two tags S0 Sam. A few indeclinable prepositions can be followed by plain personal pronouns Pbpn{m,o,y,e,i,s,u,a}w. .pp The system in older Manx is rather different: most indeclinable prepositions get person-number from a possessive pronoun Pbpp{m,o,y,n}. The tags here should be adequate, though Sa{m,o,y,i,s,u,a} will not be needed. .sh 1 "Conjunction .TS box,center; l l. POS function _ C s=subordinating c=co-ordinating .TE .fe Subordinating conjunctions Cs are .i "my, manney, mannagh, dy, na, nagh" . .fe Co-ordinating conjunctions (Cc) are .i "as, ny" . .pp Prepositions with clausal arguments, as .i dys in .i "V'eh troggit ayns Skylley Creest .bi dys .i "v'eh kaire bleeaney jeig should be tagged Si0 (S0 for an indeclinable preposition). .pp There seems to be no good reason to class .i agh , `but, except', as anything other than a preposition, tag S0. .bp .sh 1 "Numeral .TS box,center; l l l l. POS mutation type syntax _ M b=base n=nominal u=1\-9 l=lenition a=adjectival t=bound 10 o=ordinal 1=free 10 2=twenty 3=twenties f=figures .TE The first three numbers have distinct nominal and adjectival forms: .fe nominal numbers are (forms Mbnu) .i "nane, jees, troor/trass, kaire, ..." .fe adjectival numbers are (forms Mbau) .i "un, daa, three, kaire, ..." . The .fe ordinal tags apply only to words which are morphologically ordinal, e.g. .i "threeoo jeig `13th' is tagged Mbou Mbnt, but .i "three jeigoo `13th' is tagged Mbnu Mbot. .pp The .b syntax feature accounts for the different combinatory possibilities of the numbers: .in .ti 0 .fe Units M{b,l}{n,a,o}u are 1-9, .ti 0 .fe bound ten M{b,l}{n,a,o}t is .i "jeig, yeig, jeigoo, yeigoo" , .ti 0 .fe free ten Mb{n,a}1 is .i "jeih" , .ti 0 .fe twenty is .i feed (Mb{n,a}2) etc., .ti 0 .fe twenties are one-word multiples of 20, including .i ceead (Mb{n,a}3). .ti 0 A number written in .fe figures should be tagged as though it was spelt out if it would be read as a single word, e.g. .i "8, 20, 100" ; the tag M{b,l}{n,a,o}f is for numbers like .i "23, 84, 123" , which would be read as phrases, with a noun phrase infixed when used adjectivally. .sh 1 "Interjection .TS box,center; l. POS _ I .TE For example .i "she, aw, yagh" . .i She , which is used only as a one-word answer, is tagged I on the model of .i yes in the English tagset. .sh 1 "Unique .TS box,center; l l. POS mutation _ U b=base l=lenition .TE This class is inserted in order to conform to the guidelines and to allow for oversight. .sh 1 "Residual .TS box,center; l l l l l. POS mutation type _ X b=base f=foreign l=lenition s=symbol b=abbreviation .TE .sh 1 "Punctuation .TS box,center; l l l l l. POS type _ F e=sentence-final i=sentence-internal a=opening z=closing b=dash .TE Written Manx uses English-style punctuation. .sh 1 "Example Finally, an example of the tags in use: the following tagged text is from a transcription of John Kneen's account of a trip to Douglas. .ls 2 .de wo 'ie ( \w'\E$1' > \w'\E$2' ) .nr a \w'\E$1' 'el .nr a \w'\E$2' \Z'\d\d\s-2\E$2\s0'\Z'\E$1'\h'\Enau' .. .wo Yn Tg .wo traa Nbc .wo va Vib0 .wo mee Pbpnmw .wo aag Abop .wo , Fi .wo va Via0 .wo mee Pbpnmw .wo beaghey Nbv .wo ayns Si0 .wo boayll Nbc .wo na Cs .wo row Vdb0 .wo monney Abop .wo sleih Nbc .wo mygeayt Rc .wo ec Si0 .wo y Tg .wo traa Nbc .wo ; Fi .wo as Cc .wo va Via0 .wo 'n Tg .wo jishig Nbc .wo ginsh Nbv .wo duin Sis .wo paitchyn Nbc .wo \(em Fb .wo daa Mbau .wo phaitchey Nlc .wo , Fi .wo neen Nbc .wo as Cc .wo my Pbppm .wo hene Pbx .wo \(em Fb .wo v' Via0 .wo eh Pbpnew .wo goll Nbv .wo un Mbau .wo laa Nbc .wo dy S0 .wo cur Nbv .wo shilley Nbc .wo dys Si0 .wo yn Tg .wo sleidjyn Nbc .wo . Fe .wo As Cc .wo ren Via0 .wo y Tg .wo laa Nbc .wo cheet Nbv .wo as Cc .wo va Via0 .wo shin Pbpnsw .wo goll Nbv .wo dys Si0 .wo Doolish Nbp .wo . Fe .wo Ren Via0 .wo shin Pbpnsw .wo girree Nbv .wo ayns Si0 .wo y Tg .wo vorree Nlc .wo as Cc .wo ren Via0 .wo shin Pbpnsw .wo geddyn Nbv .wo fo-raad Rc .wo . Fe