The Swedish PAROLE Lexicon

A language engineering resource with access to morphological and syntactic information in Swedish,
elaborated by Språkdata, Göteborgs Universitet

Contact person Maria Toporowska Gronostaj

 

1. General design information

The Swedish PAROLE lexicon is a language engineering resource which provides access to the lexically determined morphological and syntactic information on 20000 Swedish words. It has been elaborated by Språkdata, Göteborgs Universitet. In the work of compiling the Swedish PAROLE lexicon, machine readable dictionaries such as Svenska ord - med uttal och förklaringar  (1992), and Nationalencycklopedins ordbok  (1995), and the lexical data base Göteborgs Lexikaliska Databas  (GLDB), were used as sources of lexical information on the morphological, syntactic and semantic properties of Swedish words.

The Swedish PAROLE lexicon is based on a conceptual and formal model formulated within the LE-PAROLE project, the main goal of which was to generate a common set of harmonized and reusable lexicon resources for the 12 EU languages. The Swedish lexicon is just one in a series of multifunctional lexicon resources elaborated for Catalan, Danish, Dutch, English, Finnish, French, German, Greek, Italian, Portuguese and Spanish. All those lexicons aim at supporting NLP tasks such as parsing, information retrieval and machine translation. As these tasks also require access to semantic information, the lexical data in the PAROLE lexicons is being integrated with the semantic data in the SIMPLE lexicons. (See the document on the Swedish SIMPLE lexicon)

The Swedish PAROLE lexicon and its sister lexicons are built according to the Entity-Attribute-Relation representation model. The exchange format for all the lexicons is SGML DTD. An SGML DTD (Document Type Definition) is a physical model of grammar that describes the marking of data. The DTD defines the formal structure of elements in the morphological, syntactic and semantic layers, their features and coocurrence restrictions as well as the types of relations holding between the layers. The three layers function as independent modules in the PAROLE model, but they are linked to each other through morphological, syntactic and semantic units, these being the access points to the respective layers.

A morphological unit is linked to one or more syntactic units which share the same morphological information. A syntactic unit is linked to one or more semantic units which share the same syntactic information. While the information on morphological and syntactic units and the relations between the morphological and syntactic layers is encoded in the PAROLE lexicon, the information on semantic units and the relations between syntactic and semantic layers is being encoded in the SIMPLE lexicon.

 

2. The Swedish PAROLE lexicon

In this section, an overwiev of the lexical data encoded on the morphological and syntactic layer is attempted. The syntactic layer is considered in more detail, as it plays a significant role in establishing the correspondences between syntactic and semantic layers of the lexicon model.

 

2.1 Morphological layer

An overall conceptual model for the morphosyntactic description of Swedish has been presented in the PAROLE report (WP-4.2.2b) Morphosyntactic Description of Swedish  (Danielsson, Järborg, (1996)). The morphosyntactic categories and subcategories distinguished there have been considered as a basis for the encoding morphosyntactic data on the morphological and syntactic layers.

Swedish morphological units (Um_S) and their inflectional forms have been put in focus for the description of morphological data in the Swedish lexicon. The statistics in the table below provides information on the size of the Swedish morphological lexicon seen from the point of view of the number of the encoded morphological units and the number of the morphological modes.

  Number of morphological units19985  
  Number of simple inflection modes259  

 

The distribution of the morphological units among the parts of speech is illustrated below.

  Category  SubcategoryNumber of Units  
  Adjective  none5326  
  Adposition  preposition80  
  Adverb  none450  
  Adverb  particle32  
  Article  definite1  
  Article  indefinite1  
  Noun  common11451  
  Verb  normal2604  
  Conjunction  coordinating8  
  Conjunction  subordinating30  
  Pronoun  none1  
  Unique (att- infinitive mark)  none1  

Morphological modes in the lexicon capture the types of inflectional relations a morphological unit can display. The descriptions of those morphological modes vary according to their part of speech and the idiosyncratic morphological properties associated to the morphosyntactic type. To account for these, a technical stem and a set of relevant affixes are postulated for each morphological mode.

The excerpt from Swedish morphological lexicon illustrates the inflection mode for verbs which inflect like UM_S säga  ('to say').

        <UMG NIEME="0" MF="VB_742_SAGA">
        <LIB>försäger</LIB>
        <RADG NIEME="1">
            <LIB>försade</LIB>
        </RADG>
    </UMG>
    </UM_S>
    <MFG ID="VB_742_SAGA"
        COMMENTAIRE="Verbs inflected like säga"
        EXEMPLE="säga, sade, sagt">
        <COMBTM_CFF COMBTM="PRA">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>äger</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PRP">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>ägs</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAA">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>s</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="INF">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>äga</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="IMP">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>äg</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="SUP">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>agt</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAPART">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>agd</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PEP">
            <CFF NIEME_RADGP="1">
                <RETRAIT>ade</RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>ägande</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
    </MFG>
    <COMBTM ID="PRA" TENSE="PRESENT" VOICE="ACTIVE">
    <COMBTM ID="PRP" TENSE="PRESENT" VOICE="PASSIVE">
    <COMBTM ID="PAA" TENSE="PAST" VOICE="ACTIVE">
    <COMBTM ID="PAP" TENSE="PAST" VOICE="PASSIVE">
    <COMBTM ID="INF" MOOD="INFINITIVE">
    <COMBTM ID="IMP" MOOD="IMPERATIVE">
    <COMBTM ID="SUP" MOOD="SUPINO">
    <COMBTM ID="PAPART" MOOD="PASTPART">
    <COMBTM ID="PEP" MOOD="PRESPART">
</PAROLEMORPHO>

 

2.2 Syntactic layer

The conceptual description and the formal encoding of Swedish syntactic data is conformant to the LE PAROLE model and its specifications, explicated in the "Report on the Syntactic Layer" (Guimier, Ogonowski, (1998)). The report mentioned provides a synthesis of the conceptual and formal model. It includes also the SGML DTD document, the previous knowledge of which seems to be indispensable for a thorough understanding of the formal structure of the PAROLE morphological and syntactic lexicons.

A syntactic unit (Usyn) is a basic access point to the syntactic layer. In the Swedish PAROLE syntactic lexicon, there are 28882 syntactic units linked to 19985 morphological units. These syntactic units have their syntactic behaviour described by means of 438 syntactic descriptions.

  Number of syntactic units28882  
  Number of syntactic descriptions438  

 

The distribution of syntactic units by part of speech in the Swedish syntactic PAROLE lexicon is shown below.

  CategoryNumber of Units  
  Verb3048  
  Noun14500  
  Adjective10715  
  Adverb543  
  Preposition76  

 

In the Swedish syntactic lexicon following aspects of lexically determined syntactic behaviour have been accounted for:

  • subcategorization, including functions of subcategorized elements, information on their optionality, their position and morphosyntactic realisation, syntactic and semantic restrictions
  • characteristics of the lexical unit when associated to a subcategorization frame
  • control
  • pronominalization

These lexically determined information categories have been considered as criterial for distinguishing different types of subcategorizations frames in Swedish, or in other words, for postulating 438 Description objects. The information has been encoded in the SGML DTD format by means of the following objects:

  • Description (Frame) (a syntactic behaviour: Self + Construction)
  • Self (lexical unit's characteristics when inserted into Construction)
  • Construction (list of Positions (slots))
  • Position (Slot) (a complement or an element of context)
  • Syntagma (Slot's non-terminal (NTC) or terminal realization (T))
  • Typed-feature (restrictions to be added on SyntagmaNTC, SyntagmaT or on Self)

To make the syntactic content of the lexicon more perspicuous, a closer look at some language specific linguistic criteria relevant for differentiation of syntactic descriptions for nouns, verbs, adjectives and adverbs are outlined below. In Appendix I, the excerpts from the Swedish PAROLE lexicon are provided in the standardized SGML DTD format.

 

2.2.1 Nouns

The noun descriptions in the Swedish PAROLE lexicon capture information on subcategorization frames for nouns which are heads in nominal phrases. There are 192 noun descriptions distinguished in the lexicon. All of them except one, DN0, have subcategorization frames which are lexically determined by their heads. The nouns sharing the DN0 description can relatively freely take a number of facultative pre- or postmodifing attributes in conformity to the rules of the Swedish grammar. In the premodifing positions, one can find nominal genetive attribute (Pelles/hans  'Pelle's/his'), numeral and/or adjectival attribute(s) (fem nya  'five new'). In the postmodifying position, an adverb, att -clause, relative clause or prepositional attribute can occur.

Those nouns that lexically determine their subcategorization patterns are given a more itemized description code, specifying obligatory or optional occurence of complements, their function, realization and preferences on selectional restrictions. For example, the obligatory complements are part of such noun descriptions as DNNCOMPLX (aning  'a touch (of)'), DNNCOMPLavX (anhängare  'follower'), or DNNCOMPLavAX (avsaknad  'loss, want'), while the optional ones can be found in DNNCOMPLfmedX (fusion  'fusion') and DNNCOMPLfavAX (bild  'picture'). (Some decoding clues: D description; NCOMPL noun complement; f facultative; av/med  are particular realisations of prepositions; A animatum, X inanimatum).

The syntactic unit, bild  ('picture'), USYN ID="US1710_1867_1, linked to the morphological unit UM_S ID="NO1710_1867, is given the following representation in the Swedish PAROLE lexicon (The representation below has been manually structured in order to bring out the SGML DTD objects.):

bild

    CATGRAM="NOUN"
    SSCATGRAM="COMMON"
    AUTONOMIE="YES"
    USYN_L="US1710_1867 US1710_1867_1">
    <UMG NIEME="0" APPELLATION="bild" MF="NO_ENER1">
    <LIB>bild</LIB>
    <RADG NIEME="1">
        <LIB>bild</LIB>
    </RADG>
    </UMG>
</UM_S>

<USYN ID="US1710_1867_1"
    DESCRIPTION="DNNCOMPLFAVAX">
</USYN>

<DESCRIPTION
    ID="DNNCOMPLFAVAX"
    EXEMPLE="bild (av ngn/ngt)"
    APPELLATION="subst med fakult. PPavAX"
    UM_REPRESENTANTE="alster"
    SELF="SNN"
    CONSTRUCTION="CSNNP0NPREPCOMPFAVAX">

<SELF ID="SNN"
    APPELLATION="common nouns"
    INTERVCONST="INTERVCONST25">

<INTERVCONST ID="INTERVCONST25"
    SYNTAGME_T_L="NORDCOM">

<CONSTRUCTION
    ID="CSNNP0NPREPCOMPFAVAX"
    ETIQUETTESYNT="NP">
    <POSITIONINSTANTIEE_C
        RANG="0"
        POSITION_C="PPNPREPCOMPAVAX">
</CONSTRUCTION>

<POSITION_C
    ID="PPNPREPCOMPAVAX"
    FONCTION="NPREPCOMP"
    SYNTAGME_C_L="PPORDAVAX PPOBLAVPAX">

<SYNTAGME_T
    ID="NORDCOM"
    ETIQUETTESYNT="N">
    <TRAITSYNTFERME
        VALEUR="COMMON">
</SYNTAGME_T>

<SYNTAGME_NT_C
    ID="PPORDAVAX"
    ETIQUETTESYNT="PP"
    TRAIT_L="TRAIT_LEX_PREPAV">
    <TRAITSYNTFERME
        NOM="CASE"
        VALEUR="ORDINARY">
</SYNTAGME_NT_C>

<SYNTAGME_NT_C
    ID="PPOBLAVPAX"
    ETIQUETTESYNT="PP"
    TRAIT_L="TRAIT_LEX_PREPAV">
    <TRAITSYNTFERME
        VALEUR="PERSONAL">
    <TRAITSYNTFERME
        NOM="CASE"
        VALEUR="OBLIQUE">
</SYNTAGME_NT_C>

(NORDCOM noun ordinary common
NPREPCOMP - nominal prepositional phrase
PPNPREPCOMPAVAX prepositional phrase with the prep AV followed by AX
PPOBLAVPAX prepositional phrase realized by a personal pronoun in oblique case
PPORDAVAX PPOBLAVPAX prepositional phrase realized by a noun in ordinary case
SNN common noun)

The semantic representation for this noun is given in the document The Swedish SIMPLE Lexicon.

 

2.2.2 Verbs

Verb descriptions in the Swedish PAROLE lexicon account for those subcategorization frames in which verbs function as heads in active clauses. There are 177 verb descriptions distinguished in the Swedish lexicon. These descriptions cover mono-, bi- and trivalent verbs. In order to make the encoding of subcategorizations frames for verbs consistent and homogeneous, the following subclasses av verbs have been distinguished with respect to their morphosyntactic and semantic properties:

  1. simplex verbs
  2. reflexive verbs
  3. particle verbs
  4. reflexive particle verbs
  5. modal verbs
  6. copula verbs

The information on the classes i to vi can be read off from the value of the object Self. The above verb subclassification makes it also possible to trigger off relevant sets of grammatical rules for particular verb classes. For example, reflexive verbs activate a grammatical rule that accounts for a coreferential relation between the subject and the reflexive pronoun in the clause. Particle verbs in Swedish trigger off a grammatical rule which defines particle position as either free or verb bound.

The relevant information on verbs arguments is conveyed by objects Construction and Position. The function of arguments and their realisation, the arguments' place in a construction and their optionality as well as selectional preferences with respect to feature Animatum/Inanimatum have been encoded. In a few cases, information on semantic role has been provided.

An SGML representation for the description D01P11DO (avskeda  - dismiss) is given below. The description D01P11DO exemplifies a subcategorization frame for bivalent transitive verbs with the subject and object being restricted to animate nouns.

avskedar

    CATGRAM="VERB"
    AUTONOMIE="YES"
    USYN_L="US935_1006">
    <UMG NIEME="0"
        APPELLATION="avskedar"
        MF="VB_711">
    <LIB>avskedar</LIB>
    <RADG NIEME="1">
        <LIB>avskeda</LIB>
    </RADG>
    </UMG>
</UM_S>

<USYN
    ID="US935_1006"
    DESCRIPTION="D01P11DO">
</USYN>

<DESCRIPTION
    ID="D01P11DO"
    EXEMPLE="A arrestera B"
    UM_REPRESENTANTE="arrestera"
    SELF="S1SVB"
    CONSTRUCTION="CASE1P1OBJA">

<SELF ID="S1SVB"
    INTERVCONST="INTERVCONST0">

<INTERVCONST ID="INTERVCONST0"
    SYNTAGME_T_L="V1SVB">

<CONSTRUCTION
    ID="CASE1P1OBJA"
    ETIQUETTESYNT="CLAUSE"
    INSERESELF="1">
    <POSITIONINSTANTIEE_C
        RANG="0"
        OPTIONNEL="NOO"
        POSITION_C="NPSUBA">
    <POSITIONINSTANTIEE_C
        RANG="1"
        OPTIONNEL="NOO"
        POSITION_C="NPOBJA">
</CONSTRUCTION>

<POSITION_C
    ID="NPSUBA"
    APPELLATION="animate subject nominal phrase or pronoun"
    FONCTION="SUBJECT"
    ROLETH="AGENT"
    SYNTAGME_C_L="NPORDA PRONOMPA">

<POSITION_C
    ID="NPOBJA"
    APPELLATION="animate object nominal"
    REPETABLE="NO"
    FONCION="OBJECT"
    ROLETH="PATIENT"
    SYNTAGME_C_L="NPORDA PROOBLPA">

<SYNTAGME_T
    ID="V1SVB"
    ETIQUETTESYNT="V">
</SYNTAGME_T>

<SYNTAGME_T
    ID="PRONOMPA"
    ETIQUTTESYNT="PRO">
    <TRAITSYNTFERME NOM="CASE"
        VALEUR="NOMINATIVE">
    <TRAITSYNTFERME
        VALEUR="PERSONAL">
    <TRAITSYNTOUVERT
        NOM="SSEM"
        VALEUR="ANIMATUM">
</SYNTAGME_T>

<SYNTAGME_T
    ID="PROOBLPA"
    ETIQUETTESYNT="PRO">
    <TRAITSYNTFERME
        NOM="CASE"
        VALEUR="OBLIQUE">
    <TRAITSYNTFERME
        VALEUR="PERSONAL">
    <TRAITSYNTOUVERT
        NOM="SSEM"
        VALEUR="ANIMATUM">
</SYNTAGME_T>

<SYNTAGME_NT_C
    ID="NPORDA"
    ETIQUETTESYNT="NP">
    <TRAITSYNTFERME
        NOM="CASE"
        VALEUR="ORDINARY">
    <TRAITSYNTOUVERT
        NOM="SSEM"
        VALEUR="ANIMATUM">
</SYNTAGME_NT_C>


(Syntactic codes for values:
CASE1P1OBJA Construction with an animate subject followed by a simplex verb and obligatory object
S1SVB Self restricted to simplex verb
NPSUBA subject position in a construction realized by a a nominal phrase restricted to animate nouns
NPOBJA object position in a construction realized by a a nominal phrase restricted to animate nouns NPORDA realisation of a nominal syntagma by a noun in the ordinary case (nominative/accusative form)
PROOBLPA realisation of a nominal syntagma by a pronoun in the oblique case
SSEM selectional restriction of semantic type)

The semantic representation for this verb is described in the document The Swedish SIMPLE Lexicon.

For more extensive morphological and syntactic representation and exemplification of nouns and verbs see the Swedish PAROLE-SIMPLE sample lexicon.

 

2.2.3 Adjectives

Syntactic descriptions have been provided for 1715 adjectives in the Swedish PAROLE lexicons. There are 60 different adjectival syntactic descriptions used to subcategorize adjectival frames

The main division line among the types of adjectival subcategorization frames has been drawn between those that can function as heads in predicative positions and those that function as modifiers in nominal phrases. The syntactic behaviour of the adjectives in the modifier position is defined by grammatical rules and for this reason, it is less interesting from the lexical point of view. The adjectives functioning as modifiers share the subcategorization frame encoded as DAN, in which the modifying adjective is put in the left position with respect to the head noun. The majority of the adjectives occurring in this position can also be used in a subjective predicative function and thus they subsume under the subcategorization frame of the type DACVASuPr. The descriptions DAN and DACVASuPr capture the most common subcategorization frames for adjectives.

In the Swedish PAROLE lexicon, there are many adjectives that impose lexically determined restrictions on their complements when used in predicative constructions. These have been assigned to more specific subcategorization frames. For instance, the adjectives that can take an infinitive clause as their optional complement are assigned the subcategorization frame DAINFCLf (befriande ). The class of adjectives that can undergo Tough-Movement is captured by the description type DAACOMPLfforAINFCLf.

The example in the SGML-format for the adjective befriande  ('relieving') is included in the Appendix I.

 

2.2.4 Adverbs

In the Swedish LE PAROLE lexicon, there are 543 adverbs subcategorized according to the following 8 descriptions:

   a. DADVVSADVthe adverb modifies the verb (( ) aldrig )
   b. DADVVADVclm the adverb modifies the whole clause (troligen )
   c. DADVVSprtthe adverb functions as a verb particle and modifies VP (brinna upp )
   d. DADVhthe adverb functions as a head in adverbial phrase ((mycket ) fort )
   e. DADVSFmAthe adverb modifies adjectival phrase (ganska ful )
   f. DADVSFmADVthe adverb modifies adverbial phrase (någorlunda bra )
   g. DADVnmNPthe adverb modifies nouns (bara studenter )
   h. DADVfCLthe interrogative adverbs (hur )

In the Appendix I, the description DADVnmNP in which the adverb bara  ('only') modifies the noun is given.

 

3. References

Danielsson, P., Järborg, J., 1996. The PAROLE report (WP-4.2.2b) "Morphosyntactic Description of Swedish".

Guimier, E., Ogonowski, A., 1998. Report on the Syntactic Layer, ftp://parole:yaooakgn@www.erli.fr.

Nationalencycklopedins ordbok, 1995-96. Utarbetad vid Språkdata, Göteborgs Universitet. Höganäs: Bra Böcker förlag.

Svenska ord - med uttal och förklaringar. 1992. Stockholm: Nordstedts.

 

Footnotes

 

 

PAROLE, Preparatory Action for Linguistic Resources Organization for Language Engineering, is an EC funded project aiming at generating LE resources.

 

 

SIMPLE, Semantic Information for Multifunctional Plurilingual Lexica is an EC funded project with focus on the encoding of semantic data relevant for NLP tasks.

 

 

For a more exhaustive presentation of the LE-PAROLE model, the reader is advised to turn to Guimier, Ogonowski "Report on the Syntactic Layer", 1998.

 

Appendix I

The excerpt from the Swedish PAROLE lexicon in SGML DTD format for the noun bild  ('picture'), verb avskeda  ('dismiss'), adjective befriande  (relieving') and adverb bara  ('only').

 

<PAROLEMORPHO>
    <UM_S ID="NO1710_1867" CATGRAM="NOUN" SSCATGRAM="COMMON" AUTONOMIE="YES" USYN_L="US1710_1867 US1710_1867_1">
        <UMG NIEME="0" APPELLATION="bild" MF="NO_ENER1">
        <LIB>bild</LIB>
        <RADG NIEME="1">
            <LIB>bild</LIB>
        </RADG>
    </UMG>
    </UM_S>
    <MFG ID="NO_ENER1" COMMENTAIRE="Nouns inflected like abonnent" EXEMPLE="abonnent">
        <COMBTM_CFF COMBTM="USNI">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="USND">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>en</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="UPNI">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>er</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="UPND">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>erna</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="USGI">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>s</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="USGD">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>ens</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="UPGI">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>ers</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="UPGD">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>ernas</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
    </MFG>
    <COMBTM ID="UPGD" GENDER="NONNEUTER" NUMBER="PLURAL" CASE="GENITIVE" DEFIN="DEF">
    <COMBTM ID="USGI" GENDER="NONNEUTER" NUMBER="SINGULAR" CASE="GENITIVE" DEFIN="INDEF">
    <COMBTM ID="USND" GENDER="NONNEUTER" NUMBER="SINGULAR" CASE="ORDINARY" DEFIN="DEF">
    <COMBTM ID="USGD" GENDER="NONNEUTER" NUMBER="SINGULAR" CASE="GENITIVE" DEFIN="DEF">
    <COMBTM ID="UPNI" GENDER="NONNEUTER" NUMBER="PLURAL" CASE="ORDINARY" DEFIN="INDEF">
    <COMBTM ID="UPGI" GENDER="NONNEUTER" NUMBER="PLURAL" CASE="GENITIVE" DEFIN="INDEF">
    <COMBTM ID="UPND" GENDER="NONNEUTER" NUMBER="PLURAL" CASE="ORDINARY" DEFIN="DEF">
    <COMBTM ID="USNI" GENDER="NONNEUTER" NUMBER="SINGULAR" CASE="ORDINARY" DEFIN="INDEF">
</PAROLEMORPHO>
<PAROLESYNTAXE>
    <USYN ID="US1710_1867_1" DESCRIPTION="DNNCOMPLFAVAX">
    </USYN>
    <USYN ID="US1710_1867" DESCRIPTION="DNNCOMPLFPAAX">
    </USYN>
    <DESCRIPTION ID="DNNCOMPLFAVAX" EXEMPLE="alster (av ngn/ngt)" APPELLATION="subst med fakult. PPavAX" UM_REPRESENTANTE="alster" SELF="SNN" CONSTRUCTION="CSNNP0NPREPCOMPFAVAX">
    <DESCRIPTION ID="DNNCOMPLFPAAX" EXEMPLE="bild (på ngn/ngt)" APPELLATION="subst med fakult. PPpåAX" UM_REPRESENTANTE="bild" SELF="SNN" CONSTRUCTION="CSNNP0NPREPCOMPFPAAX">
    <SELF ID="SNN" APPELLATION="common nouns" INTERVCONST="INTERVCONST25">
    <INTERVCONST ID="INTERVCONST25" SYNTAGME_T_L="NORDCOM">
    <CONSTRUCTION ID="CSNNP0NPREPCOMPFAVAX" ETIQUETTESYNT="NP">
        <POSITIONINSTANTIEE_C RANG="0" POSITION_C="PPNPREPCOMPAVAX">
    </CONSTRUCTION>
    <CONSTRUCTION ID="CSNNP0NPREPCOMPFPAAX" ETIQUETTESYNT="NP">
        <POSITIONINSTANTIEE_C RANG="0" POSITION_C="PPNPREPCOMPPAAX">
    </CONSTRUCTION>
    <POSITION_C ID="PPNPREPCOMPAVAX" FONCTION="NPREPCOMP" SYNTAGME_C_L="PPORDAVAX PPOBLAVPAX">
    <POSITION_C ID="PPNPREPCOMPPAAX" FONCTION="NPREPCOMP" SYNTAGME_C_L="PPORDPAAX PPOBLPAPAX">
    <SYNTAGME_T ID="NORDCOM" ETIQUETTESYNT="N">
        <TRAITSYNTFERME VALEUR="COMMON">
    </SYNTAGME_T>
    <SYNTAGME_NT_C ID="PPORDAVAX" ETIQUETTESYNT="PP" TRAIT_L="TRAIT_LEX_PREPAV">
        <TRAITSYNTFERME NOM="CASE" VALEUR="ORDINARY">
    </SYNTAGME_NT_C>
    <SYNTAGME_NT_C ID="PPOBLAVPAX" ETIQUETTESYNT="PP" TRAIT_L="TRAIT_LEX_PREPAV">
        <TRAITSYNTFERME VALEUR="PERSONAL">
        <TRAITSYNTFERME NOM="CASE" VALEUR="OBLIQUE">
    </SYNTAGME_NT_C>
    <SYNTAGME_NT_C ID="PPORDPAAX" ETIQUETTESYNT="PP" TRAIT_L="TRAIT_LEX_PREPPA">
        <TRAITSYNTFERME NOM="CASE" VALEUR="ORDINARY">
    </SYNTAGME_NT_C>
    <SYNTAGME_NT_C ID="PPOBLPAPAX" ETIQUETTESYNT="PP" TRAIT_L="TRAIT_LEX_PREPPA">
        <TRAITSYNTFERME VALEUR="PERSONAL">
        <TRAITSYNTFERME NOM="CASE" VALEUR="OBLIQUE">
    </SYNTAGME_NT_C>
</PAROLESYNTAXE>
</PAROLE>

 


<PAROLE LANGUE="SWEDISH">
<PAROLEMORPHO>
    <UM_S ID="VB935_1006" CATGRAM="VERB" AUTONOMIE="YES" USYN_L="US935_1006">
        <UMG NIEME="0" APPELLATION="avskedar" MF="VB_711">
        <LIB>avskedar</LIB>
        <RADG NIEME="1">
            <LIB>avskeda</LIB>
        </RADG>
    </UMG>
    </UM_S>
    <MFG ID="VB_711" COMMENTAIRE="Verbs inflected like älska" EXEMPLE="älska">
        <COMBTM_CFF COMBTM="PRA">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>r</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PRP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>s</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAA">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>de</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>des</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="INF">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="IMP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="SUP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>t</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PAPART">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>d</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="PEP">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>nde</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
    </MFG>
    <COMBTM ID="PAPART" MOOD="PASTPART">
    <COMBTM ID="SUP" MOOD="SUPINO">
    <COMBTM ID="PAP" TENSE="PAST" VOICE="PASSIVE">
    <COMBTM ID="PAA" TENSE="PAST" VOICE="ACTIVE">
    <COMBTM ID="PRP" TENSE="PRESENT" VOICE="PASSIVE">
    <COMBTM ID="PRA" TENSE="PRESENT" VOICE="ACTIVE">
    <COMBTM ID="PEP" MOOD="PRESPART">
    <COMBTM ID="INF" MOOD="INFINITIVE">
    <COMBTM ID="IMP" MOOD="IMPERATIVE">
</PAROLEMORPHO>
<PAROLESYNTAXE>
    <USYN ID="US935_1006" DESCRIPTION="D01P11DO">
    </USYN>
    <DESCRIPTION ID="D01P11DO" EXEMPLE="A arresterar B" UM_REPRESENTANTE="arrestera" SELF="S1SVB" CONSTRUCTION="CASE1P1OBJA">
    <SELF ID="S1SVB" INTERVCONST="INTERVCONST0">
    <INTERVCONST ID="INTERVCONST0" SYNTAGME_T_L="V1SVB">
    <CONSTRUCTION ID="CASE1P1OBJA" ETIQUETTESYNT="CLAUSE" INSERESELF="1">
        <POSITIONINSTANTIEE_C RANG="0" OPTIONNEL="NOO" POSITION_C="NPSUBA">
        <POSITIONINSTANTIEE_C RANG="1" OPTIONNEL="NOO" POSITION_C="NPOBJA">
    </CONSTRUCTION>
    <POSITION_C ID="NPSUBA" APPELLATION="animate subjectnominal phrase or pronoun" FONCTION="SUBJECT" ROLETH="AGENT" SYNTAGME_C_L="NPORDA PRONOMPA">
    <POSITION_C ID="NPOBJA" APPELLATION="animate object nominal" REPETABLE="NO" FONCTION="OBJECT" ROLETH="PATIENT" SYNTAGME_C_L="NPORDA PROOBLPA">
    <SYNTAGME_T ID="V1SVB" ETIQUETTESYNT="V">
    </SYNTAGME_T>
    <SYNTAGME_T ID="PRONOMPA" ETIQUETTESYNT="PRO">
        <TRAITSYNTFERME NOM="CASE" VALEUR="NOMINATIVE">
        <TRAITSYNTFERME VALEUR="PERSONAL">
        <TRAITSYNTOUVERT NOM="SSEM" VALEUR="ANIMATUM">
    </SYNTAGME_T>
    <SYNTAGME_T ID="PROOBLPA" ETIQUETTESYNT="PRO">
        <TRAITSYNTFERME NOM="CASE" VALEUR="OBLIQUE">
        <TRAITSYNTFERME VALEUR="PERSONAL">
        <TRAITSYNTOUVERT NOM="SSEM" VALEUR="ANIMATUM">
    </SYNTAGME_T>
    <SYNTAGME_NT_C ID="NPORDA" ETIQUETTESYNT="NP">
        <TRAITSYNTFERME NOM="CASE" VALEUR="ORDINARY">
        <TRAITSYNTOUVERT NOM="SSEM" VALEUR="ANIMATUM">
    </SYNTAGME_NT_C>
</PAROLESYNTAXE>
</PAROLE>

 


<PAROLE LANGUE="SWEDISH">
<PAROLEMORPHO>
    <UM_S ID="ADJ1293_1390" CATGRAM="ADJECTIVE" AUTONOMIE="YES" USYN_L="US1293_1390 US1293_1390_9">
        <UMG NIEME="0" APPELLATION="befriande" MF="ADJ_ANDE">
        <LIB>befriande</LIB>
        <RADG NIEME="1">
            <LIB>befriande</LIB>
        </RADG>
    </UMG>
    </UM_S>
    <MFG ID="ADJ_ANDE" COMMENTAIRE="Adjectives inflected like alarmerande" EXEMPLE="alarmerande">
        <COMBTM_CFF COMBTM="COMBORD">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
        <COMBTM_CFF COMBTM="COMBGEN">
            <CFF NIEME_RADGP="1">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES>s</AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
    </MFG>
    <COMBTM ID="COMBGEN" CASE="GENITIVE">
    <COMBTM ID="COMBORD" CASE="ORDINARY">
</PAROLEMORPHO>
<PAROLESYNTAXE>
    <USYN ID="US1293_1390_9" DESCRIPTION="DAN">
    </USYN>
    <USYN ID="US1293_1390" DESCRIPTION="DAINFCLF">
    </USYN>
    <DESCRIPTION ID="DAN" EXEMPLE="abstrakt (konst)" APPELLATION="Adj. in left attr. position" UM_REPRESENTANTE="abstrakt" SELF="SAAT" CONSTRUCTION="CSAATP0N">
    <DESCRIPTION ID="DAINFCLF" EXEMPLE="befriande (att INF)" APPELLATION="adj. followed by INFCL" UM_REPRESENTANTE="befriande" SELF="SADJH" CONSTRUCTION="CSAP0ACOMPINFCLF">
    <SELF ID="SAAT" APPELLATION="attributive adj" INTERVCONST="INTERVCONST_ADJAT">
    <SELF ID="SADJH" APPELLATION="adj as head in AP" INTERVCONST="INTERVCONSTXZ">
    <INTERVCONST ID="INTERVCONSTXZ" SYNTAGME_T_L="ADJH">
    <INTERVCONST ID="INTERVCONST_ADJAT" SYNTAGME_T_L="ADJAT">
    <CONSTRUCTION ID="CSAP0ACOMPINFCLF" ETIQUETTESYNT="AP">
        <POSITIONINSTANTIEE_C RANG="0" POSITION_C="ACOMPINFCLF">
    </CONSTRUCTION>
    <CONSTRUCTION ID="CSAATP0N" ETIQUETTESYNT="NP">
        <POSITIONINSTANTIEE_C RANG="0" OPTIONNEL="NOO" POSITION_C="NPNHEAD">
    </CONSTRUCTION>
    <POSITION_C ID="ACOMPINFCLF" FONCTION="ACLAUSCOMP" SYNTAGME_C_L="INFCL">
    <POSITION_C ID="NPNHEAD" SYNTAGME_C_L="NORDAX">
    <SYNTAGME_T ID="NORDAX" ETIQUETTESYNT="N">
        <TRAITSYNTFERME NOM="CASE" VALEUR="ORDINARY">
    </SYNTAGME_T>
    <SYNTAGME_T ID="ADJAT" ETIQUETTESYNT="A">
    </SYNTAGME_T>
    <SYNTAGME_T ID="ADJH" ETIQUETTESYNT="A">
    </SYNTAGME_T>
    <SYNTAGME_NT_C ID="INFCL" ETIQUETTESYNT="CLAUSE" TRAIT_L="TRAIT_LEX_INFATT">
        <TRAITSYNTFERME NOM="MOOD" VALEUR="INFINITIVE">
    </SYNTAGME_NT_C>
</PAROLESYNTAXE>
</PAROLE>

 


<PAROLE LANGUE="SWEDISH">
<PAROLEMORPHO>
    <UM_S ID="ADV1150_1238" CATGRAM="ADVERB" AUTONOMIE="YES" USYN_L="US1150_1238">
        <UMG NIEME="0" APPELLATION="bara" MF="MFGEMPTY">
        <LIB>bara</LIB>
        <RADG NIEME="1">
            <LIB>bara</LIB>
        </RADG>
    </UMG>
    </UM_S>
    <MFG ID="MFGEMPTY" COMMENTAIRE="Empty Mfg" EXEMPLE="upp">
        <COMBTM_CFF COMBTM="COMBTMEMPTY">
            <CFF NIEME_RADGP="0">
                <RETRAIT></RETRAIT>
                <AJOUTAVANT></AJOUTAVANT>
                <AJOUTAPRES></AJOUTAPRES>
            </CFF>
        </COMBTM_CFF>
    </MFG>
    <COMBTM ID="COMBTMEMPTY">
</PAROLEMORPHO>
<PAROLESYNTAXE>
    <USYN ID="US1150_1238" DESCRIPTION="DADVNMNP">
    </USYN>
    <DESCRIPTION ID="DADVNMNP" EXEMPLE="bara studenter" UM_REPRESENTANTE="bara" SELF="SADVNM" CONSTRUCTION="CSFADVNMP0NP">
    <SELF ID="SADVNM" APPELLATION="adv as noun modifier" INTERVCONST="INTERVCONST07">
    <INTERVCONST ID="INTERVCONST07" SYNTAGME_T_L="ADVNM">
    <CONSTRUCTION ID="CSFADVNMP0NP" ETIQUETTESYNT="NP">
        <POSITIONINSTANTIEE_C RANG="0" OPTIONNEL="NOO" POSITION_C="NPHEAD">
    </CONSTRUCTION>
    <POSITION_C ID="NPHEAD" SYNTAGME_C_L="NPORDAX PRONOMPAX PROOBLPAX">
    <SYNTAGME_T ID="PRONOMPAX" ETIQUETTESYNT="PRO">
        <TRAITSYNTFERME NOM="CASE" VALEUR="NOMINATIVE">
        <TRAITSYNTFERME VALEUR="PERSONAL">
    </SYNTAGME_T>
    <SYNTAGME_T ID="ADVNM" ETIQUETTESYNT="ADV">
    </SYNTAGME_T>
    <SYNTAGME_T ID="PROOBLPAX" ETIQUETTESYNT="PRO">
        <TRAITSYNTFERME NOM="CASE" VALEUR="OBLIQUE">
        <TRAITSYNTFERME VALEUR="PERSONAL">
    </SYNTAGME_T>
    <SYNTAGME_NT_C ID="NPORDAX" ETIQUETTESYNT="NP">
        <TRAITSYNTFERME NOM="CASE" VALEUR="ORDINARY">
    </SYNTAGME_NT_C>
</PAROLESYNTAXE>
</PAROLE>

Appendix II - the Swedish PAROLE-SIMPLE sample lexicon

The Swedish PAROLE-SIMPLE sample lexicon covers 80 noun units and 20 verb units. The metalanguage used for the SGML representation corresponds to the French version of the DTD.