Morph Codes [updated 6/85] 



             KEY TO CODING USED IN PACKARD'S MORPHOLOGICAL ANALYSIS

 

                     Coding for CATSS Morphological Output 

 

   

       "TYPE" CODES (3 columns maximum, to identify part of speech)

   

 N = Noun (up to 3 columns)          V = Verbs (up to 3 cols.)  

   N1 = 1st declension (fem. in -H)    (I in col.3 = augmented, 

     N1A = stem ending in -A (fem.)        whatever the stem)   

     N1M = masc. with nom. in -HS      "Progressive" Stems

     N1S = stem in -H, nom. in -A (f.)   V1 = regular present     

     N1T = masc. with nom. in -AS        V2 = contracts in -EW

   N2 = 2nd declension (masc./fem.)      V3 =   "       in -AW  

     N2N = neuters (in -ON)              V4 =   "       in -OW  

   N3 = 3rd declension                   V5 = regular -MI verbs 

     N3D,E,G,H,I,K,M,N,P,R,S,T,U,V,W     V6 = -A stem -MI verbs   

       indicate categories of 3rd        V7 = -E stem  "    "   

       decl. nouns (see other sheet)     V8 = -O stem  "    "     

   N alone= indeclinable proper noun     V9 = EI)MI/ and EI)=MI   

 A = Adjective (up to 3 cols.)         Aorist Stems               

   A1 = -OS/-H/-ON pattern endings       VA = 1st aorist active   

     A1A = -OS/-A/-ON                    VB = 2nd aorist act. #1

     A1B = -OS/-OS/-ON                   VZ = 2nd   "     "   #2(irreg) 

     A1C = -OUS/-OUS/-OUN                VH, VE, VO = -H, -E,       

     A1S = nom. in -A, stem in -H          -O stem -MI verbs    

   A3 = 3rd declension patterns          VC = #1 aor.& fut.pass.(Q-type)

     A3E,H,N,U,C as for 3 decl. nouns    VD = #2  "  "  "   "   (non-Q) 

 R = Pronouns (2 Columns)                VV = labial " " "  "     

   RA = Article                          VS = dental " " "  " (+ zeta)

   RD = Demonstrative                    VQ = guttural" " " "     

   RI = Interrogative/Indefinite       Perfect Stems  

   RP = Personal/Possessive            (I in col.3 = Plupf.augm.) 

   RR = Relative                         VX = perfect active  

   RX = O(/STIS                          VM =   "     middle    

                                         VP = labial perf. midd.  

 C = Conjunction                         VT = dental  "     " (+ zeta)

 X = Particle                            VK = guttural "     "  

 I = Interjection                      Future Stems               

 M = Indeclinable Number                 VF = regular future  

 P = Preposition                           VF2 = liquid type(+ zeta)

 D = Adverb                                VF3 = E)LAU/NW type    

                                           VFX = future perfect 



         "PARSE" CODE (up to 6 columns, as needed, to parse each form) 

 

 Nouns and Pronouns (3 columns)      Verbs (to 5 cols, exc.Ptcp)        

   col.1=case: N(om) G(en) D(at)       col.1=tense: P(resent)           

     A(ccus) V(ocative)                  I(mperfect) F(ut) A(or)

   col.2=number: S(ing) D(ual) P(l)      X(Perfect) Y(Pluperfect)       

   col.3=gender: M(asc) F(em) N(eut)   col.2=voice: A(ctive)  

 Adjectives (up to 4 columns)            M(iddle) P(assive)           

   cols.1-3 as with Nouns (above)      col.3=mood: I(ndicative)       

   col.4=degree if irregular:            D(Imperative) S(ubjunct)       

     C(ompar) S(uperl)                   O(pt) N(Infin) P(Ptcp)   

                                       col.4=person: 1 2 3

                                       col.5=number as Noun col.2 

                                       cols.4-6 Ptcpl as for Noun 



 

  

UNPACKING THE LXX MORPHOLOGICAL ANALYSIS FILES 

  

As an experiment in compressing the size of files in which the 

columnar structure produces many blank spaces, the Morphological 

Analysis files have been "packed" in the following manner: 

  

The original structure of the files designated cols. 1-25 for the  

text word, 26-36 for the analysis coding, and 37 onward for the  

dictionary form, in which as many as four separate elements might  

be recorded if a word contains three prefixal forms (rarely; two 

is more frequent, e.g. I(/STHMI A)PO KATA). The packed files 

replace any blank space between the first two fields with the  

symbol "<", and any space at the end of the second field with  

">"; if the final field contains one or more prefixed forms, the 

blank space between the root form and the first prefix is  

replaced with "@". 

  

Examples of records in unpacked and packed forms follow: 

  

KAI\                     C          KAI/ 

KAI\<C>KAI/  

  

GA/R                     X          GA/R 

GA/R<X>GA/R  

  

O(                       RA  NSM    O( 

O(<RA  NSM>O(  

  

AU)TO\N                  RD  ASM    AU)TO/S  

AU)TO\N<RD  ASM>AU)TO/S  



LE/GOUSA                 V1  PAPNSF LE/GW  

LE/GOUSA<V1  PAPNSF>LE/GW  

  

E)CHLEI/FQHSAN           VVI API3P  A)LEI/FW           E)K 

E)CHLEI/FQHSAN<VVI API3P>A)LEI/FW@E)K  

  

E)CANE/TEILEN            VAI AAI3S  TE/LLW             E)K   A)NA  

E)CANE/TEILEN<VAI AAI3S>TE/LLW@E)K   A)NA  

  

A program to unpack the MORPH files is available for IBYCUS 

and IBM/DOS machines. It reconstructs the column structure so  

that the text appears in columns 1-25, the analysis in 26-36, the  

dictionary form from 37 onward, and any prefixes in 56 onward  

(allowing 6 spaces for each prefix). Its basic algorithm is as 

follows: 

  

Get each record from the file. 

If the record begins with tilde (~), keep it intact; 

otherwise, 

if the record contains "<" and ">",  

pad the material before the "<" to 25, 

pad the material between "<" and ">" to 11,  

and eliminate the "<" and ">" symbols. 

If the record contains "@",  

pad the material after ">" and before "@" to 19, eliminate the "@",  

and retain the remaining material.