{
    "paper_id": "W89-0129",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T03:44:37.237424Z"
    },
    "title": "Lemmatising the Definitions of Svensk Ordbok by Morphological and Syntactic Analysis. A Pilot Study Lemmatizing Dictionary Definitions",
    "authors": [
        {
            "first": "A",
            "middle": [],
            "last": "Nna",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "S\u00e5gvall",
            "middle": [],
            "last": "Hein",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "In this paper we piresent the results of a study of the definition vocabu lary o f Svensk ordbok. It is part o f our ongoing work on the generation of a machine-tractable dictionary from this dictionary, in specific, of making its definitions exploitable to a parser. Aiming, in particular, at the auto matic lemmatisation of the definition vocabulary, the study includes an automatic morphological analysis o f a subset o f it, and an examination o f its results. Two major issues were addressed, i.e. the coverage of the dictionary in relation to the definition vocabulary, eind the feasibility of homograph resolution by syntactic analysis.",
    "pdf_parse": {
        "paper_id": "W89-0129",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "In this paper we piresent the results of a study of the definition vocabu lary o f Svensk ordbok. It is part o f our ongoing work on the generation of a machine-tractable dictionary from this dictionary, in specific, of making its definitions exploitable to a parser. Aiming, in particular, at the auto matic lemmatisation of the definition vocabulary, the study includes an automatic morphological analysis o f a subset o f it, and an examination o f its results. Two major issues were addressed, i.e. the coverage of the dictionary in relation to the definition vocabulary, eind the feasibility of homograph resolution by syntactic analysis.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "In th e sem a n tic m a tu ra l tr a n s fo r m a tio n ). G iv e n th is re s tr ic tio n , th e d efin i tion sh o u ld d e s c r ib e th e kernel sen se in an a n a ly tic w ay, i.e. w ith ea ch se m a n tic fa c t o r b e in g re c o g n iz e d d u rin g th e a n a ly sis b e in g a ssign ed a se p a ra te By definition vocabulary, as opposed to defining vocabulary we simply understand the inventory of words actually used in the formulation of the definitions. A s s h o w n in T a b le 1, a sin g le a n a ly sis w as g iv en in 73 p e rcen t o f th e cases, n o a n a ly sis in 23 p e rce n t o f th e ca ses, a n d , finally, m o r e th a n o n e a n a lysis fo r \u00e2",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Definitions o f S ven sk Ordbok",
                "sec_num": "2"
            },
            {
                "text": "The information on dictionary stem is saved to be used as a tool in extreu:ting a subdic tionary for the syntactic analysis to follow; the information on inflectional type is included to provide a basis for subsequent statistical calculations. m in o r ity (5 p e rce n t) o f th e (g r a p h ic ) w o rd s. B e lo w w e w ill e x a m in e th e differen t ca ses, sta rtin g w ith a n a ly sis failu re. T w o w o r d s w e re n o t a n a ly sed d u e t o lim ita tio n s in th e in fle ctio n a l g ra m m a r.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The M orphological Analysis",
                "sec_num": "3"
            },
            {
                "text": "T h e y rep resen t a ty p e th a t c a n b e illu stra te d b y th e g ra p h ic w o rd a n sikts- A b -A a n tid em ok ra tisk ' a n tid e m o c r a tic ' 2 < 1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Lexical G aps",
                "sec_num": "3.1"
            },
            {
                "text": "A -A allm \u00e4nku lturella 'g en era l c u ltu r a l' 1 < 1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Compounds and Derivatives",
                "sec_num": "3.1.1"
            },
            {
                "text": "T o t a l 4 5 2 1 0 0 % In T a b le 4 w e p resen t th e d is tr ib u tio n o f th e d eriv a tiv es b y d ifferen t ty p es.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Compounds and Derivatives",
                "sec_num": "3.1.1"
            },
            {
                "text": "T h e to t a l n u m b e r o f d eriv a tiv e s g iv en in T a b le 4 (9 0 ), is h ig h er th an th a t ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Compounds and Derivatives",
                "sec_num": "3.1.1"
            },
            {
                "text": "In T y p e A b solu te R elative frequen cy frequency",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": ".2 H om ography",
                "sec_num": "3"
            },
            {
                "text": "A / V 29 30 N / V 25 26 A / N 13 14 N 1 /N 2 10 10 V 1 / V 2 5 5 A / A b 3 3 A b / P p 3 3 N / A b 2 2 N / P n 2 2 N / I 1 < 1 V / I 1 < 1 V / P p 1 < 1 C /I m 1 < 1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": ".2 H om ography",
                "sec_num": "3"
            },
            {
                "text": "T o t a l 96 1 0 0 % In 3. A ( w e a k ) /N a m erik a n sk a ' A m e r ic a n (w o m a n )' 4",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": ".2 H om ography",
                "sec_num": "3"
            },
            {
                "text": "Total 13 banans ' b a n a n a ( g e n ) /t h e p a t h (g e n )' 1 9. N l ( g e n ) /N 2 ( g e n ) bars ' b a r l( g e n ) /b a x 2 ( g e n ) ' 1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": ".2 H om ography",
                "sec_num": "3"
            },
            {
                "text": "Total 15 In T a b le 9 a n d T a b le 10 w e p resen t th e in h e ren tly akter 'stern /acts/aft' 1 N /A b /P n allt ' (the) universe/gradually/everything' 1 A /N /A b akut 'acute/casualty department/urgently' 1 V /N l/P n an dra'state/second/other' 1 N 1/N 2/N 3 bas 'bass voice/base/thrashing' 1 T otal 16 In th o s e cases w h ere th ree o r fo u r a n a ly ses w ere g e n e ra te ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": ".2 H om ography",
                "sec_num": "3"
            },
            {
                "text": "In th is p a p e r w e h a v e p resen te d th e resu lts th a t w e a ch ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "4"
            },
            {
                "text": "Proceedings of NODALIDA 1989",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Nusvensk frekvensordbok baserad p\u00e5 tidningstext. 1. Graford. Homografkomponenter",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "All\u00e9n",
                        "suffix": ""
                    }
                ],
                "year": 1970,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "All\u00e9n, S. 1970. Nusvensk frekvensordbok baserad p\u00e5 tidningstext. 1. Graford. Homo- grafkomponenter. [Frequency dictionary o f present-day Swedish based on news paper material. 1. Graphic words. Homograph components.] Stockholm.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "The Lemma-Lexeme Model o f the Swedish Lexical Data Base",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "All\u00e9n",
                        "suffix": ""
                    }
                ],
                "year": 1981,
                "venue": "Empirical Semantics",
                "volume": "",
                "issue": "",
                "pages": "376--387",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "All\u00e9n, S. 1981. The Lemma-Lexeme Model o f the Swedish Lexical Data Base. B. Rieger [Ed.]. Empirical Semantics:376-387. Bochum.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Nusvensk frekvensordbok baserad p\u00e5 tidningstext. 4-Ordled. Betydelser",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Au\u00e9n",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Berg",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [],
                        "last": "J\u00e4rborg",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [],
                        "last": "L\u00f6fstr\u00f6m",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Ralph",
                        "suffix": ""
                    },
                    {
                        "first": "&",
                        "middle": [
                            "C"
                        ],
                        "last": "Sj\u00f6green",
                        "suffix": ""
                    }
                ],
                "year": 1980,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "AU\u00e9n, S., S. Berg, J. J\u00e4rborg, J. L\u00f6fstr\u00f6m, B. Ralph, & C. Sj\u00f6green. 1980. Nusvensk frekvensordbok baserad p\u00e5 tidningstext. 4-Ordled. Betydelser. [Frequency dictio nary o f present-day Swedish based on newspaper materird. 4. Morphemes. Mean ings.] Stockholm.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "A study o f Swedish compounds",
                "authors": [
                    {
                        "first": "O",
                        "middle": [],
                        "last": "Bl\u00e5berg",
                        "suffix": ""
                    }
                ],
                "year": 1988,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Bl\u00e5berg, O. 1988. A study o f Swedish compounds. Report No. 29, Dept, of General Linguistics. University of Ume\u00e5.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Verb-partikelkonstruktionen i svenska: syntaktiska och semantiska problem. [The verb-particle construction in Swedish: syntactic and semantic prob lems",
                "authors": [
                    {
                        "first": "E",
                        "middle": [],
                        "last": "Ejerhed",
                        "suffix": ""
                    }
                ],
                "year": 1989,
                "venue": "F\u00f6rhandlingar vid sam mankomst f\u00f6 r att dryfta fr\u00e5gor r\u00f6rande svenskans beskrivning",
                "volume": "11",
                "issue": "",
                "pages": "49--64",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ejerhed, E. 1989. Verb-partikelkonstruktionen i svenska: syntaktiska och semantiska problem. [The verb-particle construction in Swedish: syntactic and semantic prob lems.] O. Josephson, H. Strand, & M. Westman [Eds.]. F\u00f6rhandlingar vid sam mankomst f\u00f6 r att dryfta fr\u00e5gor r\u00f6rande svenskans beskrivning 11:49-64. Dept. of Nordic Languages. University of Stockholm.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "A v ' as particle and as preposition",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Hellberg",
                        "suffix": ""
                    }
                ],
                "year": 1976,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Hellberg, S. 1976. i4vsom partikel och preposition. [' A v ' as particle and as preposition.] Dept, o f Computational Linguistics. University o f G\u00f6teborg.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Towards a formalized lexicon o f Swedish. Studies in computer-aided lexicology.140-158",
                "authors": [
                    {
                        "first": "J",
                        "middle": [],
                        "last": "J\u00e4rborg",
                        "suffix": ""
                    }
                ],
                "year": 1988,
                "venue": "Almqvist Si Wiksell International. Stockholm",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "J\u00e4rborg, J. 1988. Towards a formalized lexicon o f Swedish. Studies in computer-aided lexicology.140-158. Almqvist Si Wiksell International. Stockholm. (Data linguis- tica 18).",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Forskningsprogram f\u00f6r projektet En lexikonorienterad parser f\u00f6r svenska",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "S\u00e5gvall Hein",
                        "suffix": ""
                    }
                ],
                "year": 1987,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "S\u00e5gvall Hein, A. 1987a. Forskningsprogram f\u00f6r projektet En lexikonorienterad parser f\u00f6r svenska. [Research program for the project A Lexicon-Oriented Parser for Swedish.] Dept, o f Computational Linguistics. University of G\u00f6teborg.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Parsing by means of Uppsala Chart Processor (U CP)",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "S\u00e5gvall Hein",
                        "suffix": ""
                    }
                ],
                "year": 1987,
                "venue": "Natural language parsing sj/stems",
                "volume": "",
                "issue": "",
                "pages": "202--266",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "S\u00e5gvall Hein, A. 1987b. Parsing by means of Uppsala Chart Processor (U CP). L. Bole [Ed.]. Natural language parsing sj/stems:202-266. Springer Verlag. Berlin Sc Hei delberg.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Towards a comprehensive Swedish parsing dictionary. Studies in computer-aided lexicology.268-294",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "S\u00e5gvall Hein",
                        "suffix": ""
                    }
                ],
                "year": 1988,
                "venue": "Almqvist Sc Wiksell International. Stockholm",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "S\u00e5gvall Hein, A. 1988. Towards a comprehensive Swedish parsing dictionary. Studies in computer-aided lexicology.268-294. Almqvist Sc Wiksell International. Stockholm. (Data linguistica 18).",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "The LPS inflectional grammar. A listing o f the rules. Dept, of Computational Linguistics",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "S\u00e5gvall Hein",
                        "suffix": ""
                    }
                ],
                "year": 1989,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "S\u00e5gvall Hein, A. 1989. The LPS inflectional grammar. A listing o f the rules. Dept, of Computational Linguistics. University o f G\u00f6teborg.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Creating a dictiou2iry from a lexical database",
                "authors": [
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Sj\u00f6green",
                        "suffix": ""
                    }
                ],
                "year": 1988,
                "venue": "Almqvist Sc Wiksell International. Stockholm",
                "volume": "",
                "issue": "",
                "pages": "299--338",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Sj\u00f6green, C. 1988. Creating a dictiou2iry from a lexical database. Studies in computer- aided lexicology.299-338. Almqvist Sc Wiksell International. Stockholm. (Data lin guistica 18).",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Produced at Spr\u00e5kdata ([Dept, of Computational Linguistics",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Svensk",
                        "suffix": ""
                    }
                ],
                "year": 1986,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Svensk ordbok. [A dictionary o f Swedish.) 1986. Produced at Spr\u00e5kdata ([Dept, of Computational Linguistics. University of G\u00f6teborg.] Stockholm.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Machine tractable dic tionaries as tools*and resources for natural language processing",
                "authors": [
                    {
                        "first": "Y",
                        "middle": [],
                        "last": "Wilks",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Pass",
                        "suffix": ""
                    },
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Guo",
                        "suffix": ""
                    },
                    {
                        "first": "T",
                        "middle": [],
                        "last": "Mcdonald",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Slator",
                        "suffix": ""
                    }
                ],
                "year": 1988,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wilks, Y ., D. Pass, C. Guo, T . McDonald, M. Slator. 1988. Machine tractable dic tionaries as tools*and resources for natural language processing. D. Vargha [Ed.].",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Proceedings o f the 12th International Conference on Computational Linguistics (COLING):7bO-7b5. Budapest. Center for Computational Linguistics Uppsala University Box",
                "authors": [],
                "year": null,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Proceedings o f the 12th International Conference on Computational Linguistics (COLING):7bO-7b5. Budapest. Center for Computational Linguistics Uppsala University Box 513",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "uris": null,
                "num": null,
                "text": "T h e p r im a r y le x ica l s o u r c e o f th e S w edish p a rser d e v e lo p e d in th e p r o je c t A L e x ic o n -O r ie n te d P a r s e r f o r Sw edish, L P S , (S \u00e5 g v a ll H ein 1 9 8 7 a ), is S ven sk O rd bok ' A D ic t io n a r y o f S w e d ish ' (1 9 8 6 ). It c o m p ris e s 5 8 ,5 3 6 h ead w o rd s, lem m as, re p re se n tin g 6 5 ,5 6 8 lex em es. T h e lem m a is d efin ed as a g r o u p o f w o rd form s b e lo n g in g t o th e sa m e w o r d cla ss a n d th e sa m e in fie ctio n a l p a tte rn (A ll\u00e9 n 1970; In t r o d u c t io n ). T h e d is tin c t sen ses th a t ca n b e ex p ressed b y a le m m a a re referred t o as le x e m e s (A ll\u00e9 n 1981: 3 8 2 ).S v en sk O rd bok is d ra w n fr o m a le x ica l d a ta b a se, d e v e lo p e d in th e L exical D a ta B a s e , L D B , p r o je c t (A ll\u00e9 n 1 9 7 6 ). T h u s , th e le x ica l m a te ria l is n o t on ly m a c h in e -r e a d a b le , b u t a ls o s y s te m a tic a lly o r g a n iz e d in a fie x ib le file h an d lin g s y s te m (S j\u00f6 g r e e n 1 9 8 8 ). T h e d a ta b a s e w as, h ow ev e r, p rim a rily co m p ile d fo r hu m a n u se, th u s p r e s u p p o s in g lin g u istic b a c k g r o u n d k n o w le d g e t o b e fu lly e x p lo it e d . In o th e r w o rd s , it is n o t re a d ily m a ch in e-tra cta b le in th e sen se o f th e w ord u sed b y e .g . W ilk s (1 9 8 8 ). 342 Proceedings of NODALIDA 1989, pages 342-357 T h e b a s ic se m a n tic in fo rm a tio n in S v en sk O rdbok is p r o v id e d b y th e d efi n itio n s o f th e w ord s e n s e s , ex p ressed in a su b se t o f S w ed ish . In o r d e r t o m a k e th is in fo rm a tio n u sefu l t o th e p a rser, w e h a v e t o m a k e th e im p lie d lin g u istic k n o w led g e e x p lic it, a n d t o fo rm a liz e it. A ste p to w a r d s th is g o a l is t o p a rse th e d e fin itio n s, m o r p h o lo g ic a lly a n d s y n ta ctica lly . H ere w e w ill d iscu ss th e p r o b le m s in v o lv e d in le m m a tiz in g th e (g r a p h ic ) w o r d s o f th e d e fin itio n s a u to m a tic a lly . T h e p r o c e d u r e in v o lv es tw o b a s ic step s, i.e. an a u to m a t ic m o r p h o lo g ic a l a n a ly sis, fo l low ed b y a h o m o g r a p h r e s o lu tio n p r o c e s s , o p tio n a lly ca rrie d o u t b y a u to m a t ic s y n ta c tic a n a lysis. In th e ca r r y in g th r o u g h o f su ch a sch em e, a s p e c ts o f th e im p lied lin g u istic b a ck g ro u n d k n o w le d g e w ill b e b r o u g h t u p a n d c o n c r e tiz e d . In s p e cific, th e fo llo w in g q u e stion s a re a d d resse d : \u2022 T o w h a t ex te n t a re th e w o rd s o f th e d e fin itio n s e x p lic itly d e fin ed ? \u2022 T o w h a t e x te n t are th e h o m o g ra p h ie s o f th e d e fin itio n s s o lv a b le w ith in th e c o n te x ts in w h ich th e y o c c u r ? T h e resu lts th a t w e p resen t a re b a se d o n an a u to m a t ic m o r p h o lo g ic a l a n al ysis o f a su b set o f th e d e fin itio n v o ca b u la ry .^ F ir s t, h o w e v er, w e g iv e a sh o rt d e scrip tio n o f th e d efin itio n s a n d th e u n d e rly in g p rin cip le s b e h in d th em .",
                "type_str": "figure"
            },
            "FIGREF1": {
                "uris": null,
                "num": null,
                "text": "o d e l d e v e lo p e d in th e L D B p r o je c t a n d a p p lie d in S v en sk O rd bok, e v e ry le x em e is s u p p o s e d t o h a v e a w ell id e n tifia b le a n d rela tiv e ly p re g n a n t k ern el m ea n in g , a rou n d w h ich a n u m b e r o f s u b o r d in a te , d e riv e d me^lnings are g r o u p e d ( J a r b o r g 1 988). T h e kern el sen ses o f th e le x e m e s a re d e s c r ib e d b y m ean s o f a d efin ition , a n d , o p tio n a lly , a d e fin itio n c o m p lem en t, w h erea s fo r th e d e riv e d m ea n in g s th eir rela tion s t o th e kern el m e a n in g a re s ta te d . T h e d e fin itio n \" is s u b je c t t o th e re strictio n th a t it sh o u ld b e c a p a b le o f r e p la cin g th e le x e m e in q u estion (t h e d e fin ie n d u m ) in all s y n ta c tic o r m o r p h o lo g ic a l c o n te x ts (s o m e tim e s w ith th e h e lp o f s o m e n",
                "type_str": "figure"
            },
            "FIGREF2": {
                "uris": null,
                "num": null,
                "text": "w o r d o r p h ra se in th e d e fin itio n . T h e w o rd s o f th e d e fin itio n s h o u ld , as fa r as p o s s ib le , b e m o re se m a n tica lly ce n tra l in th e le x ic o n th a n th e d e fin ie n d u m , as a first ste p tow a rd s e sta b lish in g an in h eren t d efin in g v o c a b u la r y .\" (J a r b o r g ib id .: 144 ). T h e c o r p u s o f th e d e fin ition s (d isre g a rd in g d e fin itio n co m p le m e n ts ) co m p rise s 3 6 0,144 tok en s a n d 4 3 ,1 8 4 in fle ctio n a l ty p e s (d is tin c t g r a p h ic w o r d s ). T h is is th e p rim a ry corp u s o f o u r stu d y. In th e p ilo t s tu d y r e p o r te d o n h ere, a su b set o f th e p rim a ry c o r p u s h as b e e n a n a ly sed co m p r is in g th e 2 ,5 0 0 first (in a lp h a b e tic a l o rd e r) ty p e s , fr o m abborre ' p e r c h ' t o b eh ov et ' th e n e e d ' .",
                "type_str": "figure"
            },
            "FIGREF3": {
                "uris": null,
                "num": null,
                "text": "A s a resu lt o f o u r p r e v io u s w o r k in th e L P S p r o je c t , w e d is p o s e o f a m o r p h o lo g i c a l am alyser o f S w ed ish , c o m p r is in g a ste m d ic tio n a r y co v e r in g th e h ea d w o rd s o f S v en sk O rdbok, a lo n g w ith a c o r r e s p o n d in g in fle ctio n a l g ra m m a r (S \u00e5 g v a ll H ein 1988; 1 9 8 9 ). T h e p a rsin g en g in e o f th e L P S p a rser is th e Uppsala C h art P r o c e s s o r , U C P (S \u00e5 g v a ll H ein 1 A n exam ple o f m orphological analysis T h e m o r p h o lo g ic a l d e s c r ip tio n s g e n e ra te d b y th e L P S p a rser are represen ted as D ir e c te d A c y c lic G raphs, D A G s (see e.g . S h ie b e r 1 9 8 6 ). In fo rm a tio n o n w ord cla ss, lem m a , in fle c tio n a l typ e, a n d d ic tio n a r y sterr? is a c o m p u ls o r y p a rt o f e a c h m o r p h o lo g ic a l d e s c r ip tio n . F or th e rest, th e in fo rm a tio n p r o v id e d differs w ith th e d ifferen t w o r d cla sses. F o r a n illu stra tio n , w e p resen t an e x a m p le o f a m o r p h o lo g ic a l d e s c r ip tio n g e n e ra te d b y th e m o r p h o lo g ic a l a n a ly ser (F ig . 1). T h e w o r d agat (F ig . 1) has b e e n r e c o g n iz e d as a su p in e o r pa st pairticiple form o f th e v e r b aga ' flo g ' , in te r n a l h om ograph y, o r as th e n o u n agat 'a g a te ' , e x tern a l h om og ra p h y. H o w e v e r, w ith th e a u to m a t ic le m m a tisa tio n as th e p rim a ry g o a l o f th e cu rre n t stu d y , in tern a l h o m o g r a p h ie s w ill, fo r th e tim e b e in g , b e d isreg a rd ed . In o th e r w o r d s , h om o g ra p h y in th e fo llo w in g p re se n ta tio n sh o u ld b e u n d e rs to o d as e x te r n a l h om ograph y. L ik ew ise, w h en w e p resen t d a ta o n n u m b e r o f an alyses b e lo w (T a b le 1 ), w e d isre g a rd ca ses o f in tern a l h o m o g ra p h y .",
                "type_str": "figure"
            },
            "FIGREF4": {
                "uris": null,
                "num": null,
                "text": "resulting from th e in flec tio n a l m orphological analysis.",
                "type_str": "figure"
            },
            "FIGREF5": {
                "uris": null,
                "num": null,
                "text": "h e an alysis failu res are d u e to m issin g stem s in th e d ic tio n a r y , e x c lu d in g tw o ca ses o f in su fficien t g ra m m a r co v e ra g e , a n d o n e ca se o f an u n foreseen use o f p a ren th eses (T a b le 2",
                "type_str": "figure"
            },
            "FIGREF6": {
                "uris": null,
                "num": null,
                "text": "o cu s in g fo r a m o m e n t o n th e m issin g stem s, w e h a v e t o s ta te , th a t s o m e 2 3 % o f th e w o rd s in th e d efin itio n s are n o t th e m se lv es h ea d w o rd s in th e d ic tio n a r y , a n d , co n se q u e n tly , n o t e x p lic itly d efin ed . T h e y a re, b a sica lly , o f tw o k in d s, i.e. p r o p e r n o u n s a n d d erived w ords (c o m p o u n d s a n d d e riv a tiv e s). T h e v e ry fa c t th a t th ere are d e riv e d w o rd s m issin g in th e d ic t io n a r y sh o u ld n o t su rp rise us (even if th e h igh n u m b e r o f th em d id ). A \" c o m p le te \" d ic t io n a r y is th e o r e tic a lly im p o ssib le , d u e t o , a m o n g o th e r th in g s, th e rich p o te n tia l fo r w o r d fo r m a tio n , in sp e cific, a c c id e n ta l co m p o u n d in g . \" A v s a m m a sk \u00e4 l [u trym m e] ges in te d e n tr iv ia la b e ty d e ls e n h o s en ra d sa m m a n s\u00e4 ttn in g a r, a v le d n in g a r o c h p a rtik elv e rb . D e n n a b e ty d e ls e t\u00e4 ck s in d ire k t a v o r d b o k e n g e n o m b e st\u00e5 n d sd e la rn a s d e fin itio n e r .\" 'F o r th e sa m e rea so n [space] th e trivia l m ea n in g o f a n u m b e r o f c o m p o u n d s , d e riv a tiv es, a n d p a r tic le v e r b s is n o t given . T h is m e a n in g is in d ir e c tly co v e r e d b y th e d e fin itio n s o f th e c o n s titu e n t p a r ts .' {S v en sk Ordbok 1986; P r e fa c e ). W h a t ca n b e o b je c t e d t o , h o w e v e r, is th e u se o f su ch w o r d s in th e d efin ition s. \" V id a r e h a r en str\u00e4 v a n varit a tt h \u00e5 lla an ta let o r d s o m a n v \u00e4 n d s i d e fin itio n e rn a re la tiv t lite t o c h a tt v \u00e4 lja s\u00e5 en k la ord s o m o m s t\u00e4 n d ig h e te r n a m e d g e r .\" 'F u rth er th e aim has b e e n t o k eep th e n u m b er o f w o rd s u sed in th e d e fin itio n s rela tiv ely sm a ll a n d t o c h o o s e as s im p le w ord s as th e c ir c u m s ta n c e s p e r m it .' (ib id .). H ere w e w o n ' t d iscu ss th is issue fu rth e r, b u t ra th e r e x a m in e th e d e riv e d w o rd s u sed in th e d e fin itio n s t o see i f th e y keep up w ith th e tra n sp a ren c y claim , i.e. i f th eir m e a n in g is d e riv a b le fr o m th e defin ition s o f th eir c o n s titu e n t p a rts b y m ea n s o f a set o f g en era l w o r d fo r m a tio n rules. T h e ru les o n w h ich th o s e d e riv a tio n s are b a se d , are p a rt o f th e b a ck g ro u n d k n ow l e d g e im p lie d b y th e le x ic o g r a p h e r s , a n d as su ch o f p rim a ry in terest t o us. T h u s, i f th e y fu llfil th e tr a n s p a r e n c y cla im , th e y s h o u ld b e fo rm a lize d a n d in clu d ed in th e w o r d fo r m a tio n c o m p o n e n t o f th e L P S m a ch in e d ic t io n a r y as its first p iece. A su b set o f th e im p lic itly d efin e d w o rd s o f th e d e fin ition s are in clu d e d in th e d ic t io n a r y as m o r p h o lo g ic a l e x a m p le s . T h e v e r y e x is te n ce o f th e d eriv e d w o rd s is co n fir m e d b y th e e x a m p le s , b u t still, n o th in g is sa id a b o u t th eir use a n d m ean in g; th e y are fo u n d t o b e tra n sp a re n t. In a ll, th e re are 3 ,9 3 4 m o r p h o lo g ic a l ex a m p les. O n c e th e w o r d fo r m a tio n ru les b e h in d th e d e riv e d w o rd s o f th e d e fin ition s have b e e n fo r m u la te d , th e to t a l o f th e m o r p h o lo g ic a l e x a m p le s m ig h t b e used as a test c o r p u s fo r th e re su ltin g w o r d fo r m a tio n c o m p o n e n t. T h e m issin g p r o p e r n o u n s m a k e u p a m u ch sm a ller n u m b er (3 6 ) th an th e d e r iv e d w o rd s (5 3 2 ). T h ir ty -t h r e e o f th e m n a m e g e o g ra p h ica l u n its, a n d th e re m a in in g n a m e p e rso n s. A s lo n g as th e w o rd s th a t th e y d efin e are in clu d ed in th e d ic tio n a r y , w e see n o o th e r so lu tio n th a n t o en ter th e m as h ea d w ord s an d th u s d efin e th e m . A fre q u e n t ty p e is m a d e u p o f c o u n t r y n a m es (e .g . A lg e r ie t ' A lg e r ia ' ) u sed in th e d e fin itio n s o f p e o p le fr o m th o s e co u n trie s (e .g . a lg erier ' A lg e r ia n ' : p e r s o n fr \u00e5 n A lg e r ie t ' p e rs o n fr o m A lg e r ia ' ).",
                "type_str": "figure"
            },
            "FIGREF7": {
                "uris": null,
                "num": null,
                "text": "'fa c e ' in c o o r d in a t e d c o m p o u n d s su ch as a n sik ts-och h \u00e5rv\u00e5rd ' fa ce a n d h air c a r e ' . T h e p h e n o m e n o n is reg u la r a n d h as t o b e ta k en ca re o f p a r tly b y th e m o r p h o lo g i ca l, p a r tly b y th e s y n ta c tic g ra m m a r. In th e first in sta n ce , th e L P S in fle ction a l g r a m m a r w ill b e a c c o r d in g ly e x te n d e d . S h ort fo r beford rats ' (h a s b e e n ) s e n t' o r befordras ' is (b e in g ) s e n t' w e find b e fo r d r a (t)s a m o n g th e d e fin itio n w o r d s . T h is a b b r e v ia te d fo r m o f exp ressin g a lte rn a tiv e s w as n o t foreseen b y th e L P S in fle ctio n a l g ra m m a r. N o r w ill it be, b u t th e sh o rt fo r m s o f th is k in d w ill b e sp e lle d o u t.",
                "type_str": "figure"
            },
            "FIGREF8": {
                "uris": null,
                "num": null,
                "text": "T o keep u p w ith th e tra n sp a ren cy claim , th e w o r d fo r m a tio n c o m p e te n c e on w h ich th e u n d e rs ta n d in g o f (s o m e p a rts o f ) th e d e fin itio n s is b a sed , sh ou ld be m o r e s te r e o ty p ic a l th a n w h a t is g e n e ra lly th e ca se. F u rth e r, w e m u st requ ire th a t th e c o n s titu e n t w ord s are th e m s e lv e s defined, a n d , th a t th e y cire n o t h o m o g ra p h ic o r p o ly s e m o u s . W e b e g in o u r e x a m in a tio n w ith th e m o st d o m in a n t w ord T y p e E x a m p le A b s o l u t e f r e q u e n c y R e la t i v e f r e q u e n c y N -N a ktie\u00e4 ga re 's to c k h o ld e r ' 376 83 N -A P abborrliknande ' resem b lin g a p e r c h ' 22 5 N -P P a rvsber\u00e4 ttiga d 'e n titled t o an in h e r ita n c e ' 11 2 N -A alkoh olfri ' n o n -a lk o h o lic ' 18 4 A b -N bak\u00e5tspark ' k ick b a c k w a r d s ' 8 2 A -N allm \u00e4 n giltigh et ' u n iversal a p p lic a b ility ' 4 < 1 A b -A b a k terifr\u00e5 n 'fr o m th e s te r n bak\u00e5triktad ' p o in tin g b a c k w a r d s ' 3 < 1",
                "type_str": "figure"
            },
            "FIGREF9": {
                "uris": null,
                "num": null,
                "text": "fo r m a tio n ty p e , i.e. th e com pou n d s. In T a b le 3 w e p resen t th e d is tr ib u tio n o f th e co m p o u n d s b y s y n ta c tic types.\u00ae A m o n g th e 11 c o m p o u n d ty p e s th a t w ere fo u n d , th e q u ite d o m in a t in g o n e is th e N -N t y p e m a k in g u p s o m e 8 3 p e rce n t o f th e to ta l n u m b e r o f c o m p o u n d s . T h e d o m in a n ce o f th is ty p e is in a c c o r d a n c e w ith o u r in tu itio n a n d w ith o th e r stu d ies o f m o d e rn S w edish (se e e .g . B l\u00e5 b e r g 1 9 8 8 ). O u r fig u re is, h ow ev er, c o n s id e r a b ly h igh er th an th a t p resen te d b y B l\u00e5 b e r g (6 8 % ). H is figu res a re b aaed o n an in ves tig a tio n o f a co rp u s o f 3 ,9 7 1 c o m p o u n d s , id en tified in n e w sp a p e r te x t fr o m 1985. O u r h ig h er p r o p o r tio n o f N -N com p o u n d s seem s t o s u p p o r t th e tra n s p a re n cy cla im , in d ic a tin g a re stricte d use o f th e v a ria tio n o ffe re d b y th e w o r d fo r m a tio n p o te n tia l. A d d itio n a l s u p p o r t is g iv en b y th e p a ra m e te r n u m b er o f d ifferen t types, w h o s e va lu e is m u ch low er in o u r m a teria l th a n in th a t o f B l\u00e5 b e r g , i.e. nine^ versu s 22. O u r ty p e s co n s titu te a s u b se t o f th e set id e n tifie d b y B l\u00e5 b e r g , in clu d in g , in a d d itio n : V -N , V -A , A -V , A b -V , N u m e r a l-V , N u m e ra l-N , N u m e r a l-A , N u m e ra l-N u m era l, N -P r o p r iu m , A -P r o p r iu m , P r o p r iu m -N , P r o p r iu m -V , P r o priu m -P ro p riu m . It rem ain s t o b e seen th o u g h , h o w sta b le o u r fig u res rem a in th ro u g h o u t th e m a teria l. W e d o n 't, h ow e v er, e x p e c t t o fin d su ch a rich variar tio n o f ty p e s in v o lv in g p r o p e r n ou n s as d o e s B l\u00e5 b e r g , p r o p e r n o u n s in g e n e ra l, b e in g o u ts id e th e s c o p e o f S v en sk Ordbok. A m o n g th e 3 7 6 N -N c o m p o u n d s , th ere a re in a ll 154 differen t first c o n s tit uents (s im p le x o r d e r iv e d ). T h e m a jo r ity o f th e m (1 4 6 ) a re e x p lic itly d e fin e d , seven o f th em im p licitly , a n d o n e o f th e m {A 4 in th e c o m p o u n d A 4 -fo r m a t n o t at a ll). A m o n g th o s e th a t are o n ly im p lic itly d e fin e d , th ere is a d o m in a tin g ty p e , i.e. p ro ce s s n o u n s d eriv e d fr o m v e rb s b y m e a n s o f th e -( n ) in g su ffix, su ch as a v fy m in g s -' firin g ' (c f. a v fy m in g s m e k a n is m ' firin g m e ch a n is m ' ) o f th e v e rb AP is short for active present participle, and PP for passive past participle. *This is the figure we arrive at if we adapt to Blaberg, who includes the active and the passive participles in the verb group. a vfyra ' fire ' . It a c c o u n ts fo r fiv e o f th e seven cases. O n e o f th e rem a in in g cases is in it s e lf an N -N c o m p o u n d , i.e. bastuba ' bass t u b a ' (cf. b a stu b ein stru m en t ' bass tu b a in s tr u m e n t' ). T h e o th e r o n e is an A -N c o m p o u n d , i.e. andraklass ' secon d c la s s ' as in a n d ra k la ssu trym m e ' s e c o n d cla ss a r e a ' . W h e n o n e o f th e n ou n s o f an N -N c o m p o u n d is in it s e lf o n ly im p lic itly d e fin e d , th e d e riv a tio n o f th e defin ition w o r d h as t o p r o c e e d in tw o step s. W e th in k th a t su ch a situ a tio n sh ou ld b e a v o id e d , <ind th e tw o c o n s titu e n ts o f a c o m p o u n d u sed in a d efin ition b o th b e ex p licitly d efin ed . A n o t h e r p r o b le m w ith th e d e riv a tio n o f m e a n in g fr o m th e N -N c o m p o u n d s o f th e d e fin itio n s c o n c e r n s h o m o g r a p h y a n d p oly sem y . A m o n g th e first co n s titu e n ts o f th e N -N c o m p o u n d s w e fo u n d fo u r ca ses o f h o m o g ra p h y , i.e. akter, arm , babords, a n d back. A k te r , babord, a n d back a re all n o u n s o r a d v e rb s, a n d a rm is a n o u n o r an a d je c tiv e . F o r in s ta n ce , a k te r is a n ou n ' ste r n ' o r an a d v e rb ' a ft' , a n d back, a n o u n ' b a c k , r e v e rs e ' o r an a d v e r b ' b a c k ' . T h e w o rd a k te r o c c u r s in akterd \u00e4ck, a k term a st, a n d akterv\u00e4gg, a n d in d e c id in g w h e th e r th e n ou n o r the a d v e r b in te r p r e ta tio n o f a k te r w as th e in te n d e d o n e , w e w ere g u id ed b y th e m o r p h o lo g ic a l e x a m p le a k terd \u00e4 ck p resen ted u n d e r th e (n o u n ) h ea d w o rd akter. B y a n a lo g y , w e ch o s e th e n o u n in te rp re ta tio n in th e o th e r tw o ca ses, t o o . A s regards back it o c c u r s in back verk a n ' reverse e ffe c t' a n d in back\u00e5kning 'd o w n h ill g o in g ' . N o m o r p h o lo g ic a l e x a m p le s a re g iv e n , b u t in tu itiv e ly back in backverkan sh ou ld b e u n d e r s t o o d as an a d v e r b in ste a d o f as a n ou n . T h is in tu itio n is co n firm e d b y its c o n t e x t v\u00e4n d a m ed u tn y ttja n d e av backverkan ' tu rn u sin g reverse efi^ect' . T h e c o n te x t o f back\u00e5kning, i.e. (ty p a v) k\u00e4lke f \u00f6 r back\u00e5kning ' k in d o f sled g e fo r d o w n h ill g o in g ' c o n fir m s th e in u titio n th a t back h ere is a n o c c u r r e n c e o f backe ' h ill' ra th e r th a n o f back. In a d d itio n t o th e fo u r ca ses o f h o m o g r a p h y m en tion ed a b o v e , th e re a re 3 6 ca se s (9 % ) o f p o ly s e m y c o n c e r n in g th e first co n stitu e n ts o f o u r N -N m a te ria l. W it h o u t s cru tin iz in g th e in d iv id u a l ca ses fu rth e r, w e co n lu d e th a t th e d e r iv a tio n o f m e a n in g fr o m th e c o n s titu e n t p a rts o f a c o m p o u n d in th e d e fin itio n h as t o b e b a se d o n th e id e n tifica tio n o f th eir d e fin ition s. W h e n h o m o g r a p h y is in v o lv e d , a c h o ic e h as t o b e m a d e . O n ly w h en a m o r p h o lo g ic a l e x a m p le is th ere t o fa c ilita te th e c h o ic e ca n w e m a in ta in th e tra n sp a ren cy claim . T h e c o m p o u n d w o r d s w ith a n a m b ig u o u s first o r s e c o n d c o m p o n e n t w ill all have t o b e e x p lic itly d e fin e d in o u r m a ch in e d ictio n a ry . In a d d itio n t o th e 3 7 6 N -N c o m p o u n d s th ere are 76 c o m p o u n d s o f differen t s y n ta c tic s tr u c tu r e (T a b le 3 ). T h e y w ill all b e e x a m in e d w ith rega rd t o e x p licit d e fin itio n a n d h o m o g r a p h y o f c o n s titu e n t p a rts. T h e to t a l n u m b e r o f d eriv a tiv e s (n o t e x p lic itly d e fin e d ) th a t w ere iden tified is 8 0 , w h ich m ea n s th a t th e ir sh a re o f th e to t a l n u m b er o f im p lic itly defined w o r d s is o n ly r o u g h ly 15 p e r c e n t as c o m p a r e d t o 85 p e rce n t fo r th e c o m p o u n d s .",
                "type_str": "figure"
            },
            "FIGREF10": {
                "uris": null,
                "num": null,
                "text": "re se n te d in T a b le 3 (8 0 ), fo r th e fo llo w in g rea son . T h e figures in T a b le 3, are b a s e d o n a ' to p -le v e l' cla ss ifica tio n o f th e w o rd s. In o th e r w o rd s, a w o r d su ch as bakningsredskap ' b a k in g t o o l ' is cla ssified as a c o m p o u n d o n ly , d isreg a rd in g th e fa c t th a t its first c o n s titu e n t is a d eriv a tiv e n o t e x p lic itly d efin ed . H ow ev er, th e figu res p re se n te d in T a b le 4 , in c lu d e a ls o su ch in d ire ct d e riv a tio n , a n d , in so m e Types o f derivatives incl. particle verbs. cases, a w o r d co n trib u te s t o m o r e th a n o n e o f th e d e riv a tio n ty p e s . F o r in s ta n ce a w o rd su ch as avfj\u00e5llning w h ich h as t o b e d e riv e d in tw o ste p s, i.e. fj\u00e4 lla ' p e e l' (e x p lic it ly d e fin e d ), fi\u00e4lla av ' p e e l o f f ' , a n d th e p r o c e s s avfj\u00e5lln in g is c o u n t e d as an in sta n ce o f b o t h th e V P a r tic le a n d th e N { -( n ) i n g } ty p e . C a se s o f tw o -s te p d e riv a tion , th o u g h , are rare. T h e d o m in a tin g ty p e (4 9 % ) is th a t o f th e v e rb a l n o u n s, fo r m e d b y m ea n s o f th e -( n ) in g su ffix. Its 44 m em b e rs (in fle ctio n a l fo r m s ) a re d e riv e d fr o m 42 d iffer en t v e rb s. 32 o f th ese v erbs are e x p lic itly d e fin ed , w h erea s 10 o n ly im p lic itly so. F urther, s ix o f th e 32 e x p lic itly d efin ed v e rb s are p o ly s e m o u s , a n d , co n seq u en lty , th e n ou n s d e riv e d fr o m th em n o t u n iq u ely d efin ed . T h e te n im p lic itly defin ed verbs, are e ith e r p a rticle v erb s fo r m e d b y m ea n s o f th e p a r tic le av, o r p refix ed verbs fo r m e d b y m ean s o f th e p refix av. In tu itiv ely , w e id e n tify th e m as d eriv ed fro m th e p a r tic le v erb s e.g . klippa av. H ow ev er, fo r m a lly th e y m ig h t as w ell be d eriv ed fr o m th e c o rre s p o n d in g p re fix e d v e rb s e.g . avklippa. T h e sa m e k in d o f a m b ig u ity is a p r o b le m in th e id e n tific a tio n o f th e v e rb s o f p re fix e d p a st p a rtici ples, e.g . avhuggen. It m ig h t b e d e riv e d fr o m avhugga as w ell as fr o m hugga av. A s reg a rd s th e sem a n tic d is tin c tio n b e tw e e n th e p refix ed v e rb a n d th e p a rticle v erb , th e m e a n in g o f th e p refix ed v e rb seem s t o b e m o r e a b s tr a c t th a n th a t o f th e p a rticle v e rb (see fu rth er H e llb e rg 1976; E je rh e d 1 9 7 9 ). I f w e a llo w th e use o f im p lic itly d efin ed p a rticle v e rb s o r p re fix e d v e r b s as a b a sis o f v e r b a l n ou n s o f th e (n )in g ty p e , a s y ste m a tic a m b ig u ity w ill b e cr e a te d . It sh o u ld b e a v o id e d , as sh ou ld a lso th e u se o f p refix ed p a st p a rticip le s in th e d e fin itio n s, in h eren tly a m b ig u o u s as th e y are.C o n c e r n in g th e rem a in in g d e riv a tiv e ty p e s , th e ir rep resen ta tiv es w ill b e e x a m in ed fo r h o m o g ra p h ie s a n d im p lic itly d efin ed ty p e s in th e sa m e m a n n er as th e v erb a l n ou n s. In d iscu ssin g th e d eriv a tiv e s a n d th e ir a p tn ess fo r b e in g h a n d ledb y th e w o r d fo r m a tio n c o m p o n e n t ra th e r th a n b e in g re g istered as le x ic a l units, th e p r o d u c tiv ity o f th e affixes is an a d d itio n a l p a ra m e te r t o b e c o n s id e r e d . A n e x a m p le o f a sufiSx se q u e n ce w ith lo w p r o d u c t iv it y is -o r -is k w ith a n a b s o lu te (le x ic a l) firequency o f o n e (a rtik u la to risk ' a r tic u la to r y ' ) in o u r m a te ria l a n d o f tw o in th e N F O m a teria l (A ll\u00e9 n et al. 1 9 8 0 ). F in a lly , d e riv e d w o rd s, w h ich ca n b e s eg m en ted in m o r e th an o n e w a y sh ou ld b e a v o id e d . A n e x a m p le o f su ch an in sta n ce in o u r m a teria l is aU ian sfri-h et ' n o n -a lig n m e n t' o r a llia n s-jrih et.",
                "type_str": "figure"
            },
            "FIGREF11": {
                "uris": null,
                "num": null,
                "text": "th e a u to m a t ic le m m a tiz a tio n p r o c e s s , th e m o r p h o lo g ic a l a n a lysis has to b e fo llo w e d b y a h o m o g r a p h r e s o lu tio n p r o c e d u r e . T h e n eed fo r su ch a p r o ce d u re in o u r m a te ria l is e v id e n t & om th e figu res o n a lte rn a tiv e em alyses p resen ted in T a b le 1. R o u g h ly fiv e p e r c e n t o f th e d e fin itio n w o rd s are h o m o g r a p h ic . H ere we w ill g iv e a p re s e n ta tio n o f th e differen t k in d s o f h o m o g ra p h ie s th a t w e fo u n d , a n d d iscu ss th e p o s s ib ility o f s o lv in g th e m b y m ea n s o f s y n ta c tic an alysis.",
                "type_str": "figure"
            },
            "FIGREF12": {
                "uris": null,
                "num": null,
                "text": "T a b le 5 w e p re se n t a w o r d cla ss b a se d o v e rv ie w o f th e differen t k in ds o f h o m o g r a p h ie s re su ltin g fr o m th e m o r p h o lo g ic a l a n a ly sis. F irst w e sta te , th a t in r o u g h ly 85 p e rce n t o f th e ca se s, th e h o m o g r a p h c o m p o n e n ts b e lo n g t o differen t w o r d cla sses, w h erea s 15 p e r c e n t c o n c e r n h o m o g r a p h y w ith in o n e w o r d cla ss (th e n o u n s o r th e v e r b s ). T h e p r e c o n d itio n s fo r s o lv in g th e h o m o g r a p h ie s b y sy n ta c tic m e a n s, s h o u ld , o f co u r s e , b e b e tt e r w ith in th e first c a te g o r y th a n w ith in th e s e c o n d o n e , w h e re w e m a y c o m e c lo s e t o p u r e ly le x ic a l d is a m b ig u a tio n . H ow ever, fo r a n ev a lu a tio n o f th is g e n e ra l a s s u m p tio n , w e n ee d t o k n ow m o r e a b o u t th e a c tu a l fo r m s th a t c o in c id e , a n d p resen t su ch d a ta fo r th e m a jo r ty p e s in T a b les 6 t o 10. T h e o v e r w h e lm in g n u m b e r o f ca ses o f A / V h o m o g r a p h y (9 0 % ) a re d u e to c o in c id e n c e b e tw e e n th e a d je c tiv e a n d a p a r tic ip ia l fo r m o f th e v e rb . T h e d if fe r e n tia tio n b e tw e e n th e m is n o to r io u s ly d ifficu lt, t o a g re a t ex te n t d u e t o th eir p a r tly o v e r la p p in g d is tr ib u tio n . H ow e v er, o n e o f th e p rin cip le s th a t w as adh ered an sedd ' r e s p e c t e d /c o n s id e r e d ' 9 2. A ( w e a k ) / V ( p p / s u p ) ansedda ' r e s p e c t e d /c o n s id e r e d ' 3 3. A ( n e u t r ) /V ( p p /s u p ) a v a n cera t ' a d v a n c e d ' 6 4. A ( w e a k ) / V ( p p / p r e t ) ) alliera d e ' a llie d ' 6 5. A / V ( a p ) a n sl\u00e5 en d e ' im p r e s s iv e /im p r e s s in g ' 2 6. A ( w e a k ) /V ( i n f ) an rika ' h ig h -b o r n /c o n c e n t r a t e",
                "type_str": "figure"
            },
            "FIGREF13": {
                "uris": null,
                "num": null,
                "text": "t o in th e fo rm u la tio n o f th e d e fin ition s, i.e. th e r e p la c e m e n t r e s tr ic tio n (see 2 ) , m ak es th e ta sk m o re rea listic th a n w o u ld b e th e ca s e w ith u n re s tricte d te x t.T h e n in e m em b ers o f th e first ty p e o f T a b le 6 are in th e b a s ic fo r m , a n d th u s m a y o c c u p y b o th an a ttr ib u tiv e a n d a p r e d ic a tiv e p o s it io n . In a ll, th e y o c c u r 73 tim e s, a n d in 6 7 o f th ese ca ses th e a d je c tiv a l in te r p r e ta tio n sh o u ld b e p referred . O u r ch o ic e w as b a se d o n th e fo llo w in g h eu ristics:1. I f th e cu rren t w o rd is in th e a ttr ib u tiv e p o s it io n , a n d n o t n e g a ted b y an a d v e rb , c h o o s e th e a d je c tiv e . 2. I f th e cu rren t w o rd is th e h ea d o f th e d e fin itio n o f an a d je c tiv e , a n d n o t n e g a te d b y an a d v e rb , c h o o s e th e a d je c tiv e . 3. I f th e cu rren t w o rd is n o m in a liz e d , c h o o s e th e a d je c tiv e . 4. I f th e cu rren t w o r d o c c u r s in th e c o n te x t eg en sk a p en a tt vara . . . th e p r o p e rty o f b e in g c h o o s e th e a d je c tiv e . 5. E lse, c h o o s e th e p a rticip le. E xa m p les: 1. [nn a bsess]: begr\u00e4nsad an sam lin g v a r ' lim ite d a m o u n t o f p u s ' 2. [av k ry sta d ]: p \u00e5 fa lla n d e a n str\u00e4 n g d och on a tu rlig ' s trik in g ly fo r c e d a n d u n n a tu ra l' 3. [nn a rv Ij: \u00f6v erg \u00e5 n g av eg en d om av (v is s t) v\u00e4rd e f r \u00e5 n avlid en till e fte r le v a n de ' tra n sitio n o f p r o p e r ty o f (c e r ta in ) v alu e fr o m d e ce a s e d t o s u r v iv in g ' 4. [nn b e g \u00e5 v n in g /lj : eg en sk a p en a tt vara beg\u00e5vd ' th e p r o p e r t y o f b e in g ta l e n te d ' 5. [vb g \u00e4 lla 1 /3 ] : vara an sedd (s o m ) ' b e co n s id e r e d ( a s ) ' , [nn b r ic k a /1 ]: b\u00e4rbar skiva begr\u00e4nsad av l\u00e5g k a n t ' p o r ta b le p la te s u rro u n d ed b y a lo w e d g e ' . [av o a v g jo r d /1 ] : in te a vgjord till n \u00e5 g on s f\u00f6 r d e l ' n o t d e cid e d t o a n y o n e 's a d v a n ta g e ' , [nn \u00f6 d e s b y g d s v \u00e4 g ]: v\u00e4g g e n o m e j bebodda tra k ter ' roa d th rou g h n o t in h a b ite d re g io n s ' . T h e h eu ristic ru le 2 is b a se d o n th e re p la ce m e n t re strictio n ; o n ly if th e defi n itio n b e lo n g s t o th e sa m e s y n ta c tic c a te g o r y as th e d e fin ien d u m is it c a p a b le o f s u b s titu tin g it. F u rth er w e re q u ire sa m e s y n ta c tic ca teg o ry t o m ean sa m e w ord class, (o r sa m e w o r d cla ss w ith re sp e ct t o th e h ea d w o r d o f th e d e fin itio n ). If, h o w e v e r, th e h o m o g r a p h is p r e c e d e d b y a n e g a tin g a d v e rb , b e in g a s tr o n g v erb sig n a l, th e p a r tic ip le in te rp re ta tio n is ch o sen . C o n c e r n in g T y p e 2 (in T a b le 6 ) w e h a v e th e sa m e situ a tio n as w ith T y p e 1 w ith an a d d itio n a l su p in e fo r m a ltern a tiv e. T h e su p in e fo r m has a q u ite differen t d is tr ib u tio n as c o m p a r e d t o th e a d je c tiv e a n d th e p a rticip le , an d thus is easily d istin g u ish e d in th e s y n ta c tic a n a ly sis. (N o su p in e fo rm s, h ow ev er, w ere fou n d a m o n g th e 31 o c c u r r e n c e s o f th is t y p e ). In d istin g u ish in g b e tw e en th e a d jectiv a l a n d th e p a r tic ip ia l fo rm s , th e h eu ristics listed a b o v e s h o u ld b e a p p lie d . It sh ou ld a ls o w o rk o u t fo r T y p e 3 , 4 , a n d 5. T h e fin ite (p a s t te n se) fo r m o f T y p e 4 sh ou ldb e e a sily id e n tifie d in th e s y n ta c tic a n a lysis. F in a lly, T y p e 6 sh ou ld ca u se n n e u t r ) /N basalt ' b a s a l/b a s a lt ' 1",
                "type_str": "figure"
            },
            "FIGREF14": {
                "uris": null,
                "num": null,
                "text": "o m in a tin g t y p e o f th e a d je c t iv e /n o u n h o m o g r a p h y is th a t betw een th e ir b a s ic fo r m s ( T y p e 1 in T a b le 7 ). A m o n g th e 70 o c c u r r e n c e s o f th is ty p e , 4 6 a re a d je c tiv e s , a n d , o n ly 24 n ou n s. In th e a ttr ib u tiv e p o s itio n , th ere are h a r d ly a n y p r o b le m s in r e c o g n izin g th e a d je c tiv e s o n a p u re ly s y n ta c tic basis. M a k in g th e h o m o g r a p h r e so lu tio n th o u g h in th e p o s itio n a fter th e fin ite v erb re q u ire s a d d itio n a l k n o w le d g e . T h is is th e ca se, fo r in sta n ce, fo r a w o r d such as ban kru tt ' b a n k r u p c y ' in th e d e fin itio n g \u00f6ra bankrutt ' b e c o m e b a n k r u p t' [vb b a n k ru tte ra ]. F u rth e r, T y p e 2 (T a b le 7 ) is a n e x a m p le o f a ca u sa l co in c id e n c e b e tw e e n a n e u te r a d je c tiv a l fo r m a n d a n o u n . M o r e s y s te m a tic , h ow ever, is th e h o m o g r a p h y b e tw e e n th e w e a k ly in fle cte d a d je c tiv e a n d th e n ou n in T y p e 3.T h e ru les fo r n o m in a liz in g th e a d je c tiv e s m u st b e v e r y re stricte d in th e an alysis g r a m m a r , i f w e a re t o s o lv e th e se h o m o g r a p h ie s sy n ta ctica lly , th e rea son b e in g th a t th e w e a k ly in fle cte d a d je c tiv e s d u e t o th e ir in h eren t defin iten esse c o n s titu te a p r o d u c t iv e b a sis fo r n o m in a liz a tio n . (+ g en )/V (su p pass) ansats 'approach(+gen)/(been) cultivated' 1 7. N (gen)/V (pres pass) beg\u00e4rs 'desire(gen)/(is) required' 1 8. N (gen)/V (pret pass) bars 'baj(gen)/(w as) carried' 1 9. N (gen)/V (dep) andas *spirit(gen)/breathe' 1 10. N (pl)/V (pres) bakar 'backs/bake(s)' 7 Total 29 Table 8: Types o f N / V hom ographies. In T a b le 8 w e p resen t th e n o u n /v e r b h o m o g ra p h ie s . A ll th e differen t ty p e s seem to b e so lv a b le in th eir lo c a l c o n te x ts in c o m b in a t io n w ith th e s y n ta c tic p r e d ic tio n m a d e b y th e w o rd cla ss m a rk er o f th e d e iin ie n d u m , th e su b s titu tio n l ( d e f s g ) / N 2 ( d e f sg) a n d en ' th e w ild d u c k /t h e s p ir it' 1 5. N l / N 2 ( p l ) basar ' b a z a a r /b a s s v o ic e s ' 2 6. N l / N 2 ( d e f ) banan ' b \u00e5 n a n a /t h e p a t h ' 1 7. N l /N 2 ( g e n ) ask es ' a s c e tis is m /a s h w o o d ( g e n ) ' 1 8. N l ( i n d e f g e n ) /N 2 ( d e f g en )",
                "type_str": "figure"
            },
            "FIGREF15": {
                "uris": null,
                "num": null,
                "text": "m o s t d ifficu lt h o m o g r a p h y cases, i.e. th o se in w h ich th e h o m o g r a p h c o m p o n e n ts b e lo n g t o th e sa m e w o r d class, th e n ou n cla ss in T a b le 9 a n d th e v e r b cla ss in T a b le 10. A m o n g th e n o u n s th ere are fo u r ty p e s (2 , 3 , 4 , a n d 9 ) w ith n o fo r m a l cr ite r ia t o d istin g u ish b e tw e en th em . C on seq u e n tly , in th ese ca ses th e h o m o g r a p h r e s o lu tio n a m o u n ts t o p u re ly lex ica l d isa m b ig u a tio n , a n d th e h o m o g r a p h y w ill b e in d ifferen t t o th e s y n ta c tic an alysis. In T y p e 1 th e h o m o g r a p h c o m p o n e n ts d iffer w ith reg a rd t o g e n d e r on ly . Its fiv e m e m b e rs h ave, in a ll, 35 o c c u r r e n c e s , b u t o n ly in 9 o f th ese cases is g en d er d e c is iv e fo r th e c h o ic e . In T y p e 5, n u m b e r is th e d istin g u ish in g fea tu re. It has tw o m e m b e rs w ith 10 o c c u r r e n c e s . T h r e e o f th ese ca ses ca n b e rea d ily so lv e d b y s y n ta c tic m ea n s. T h r e e o f th e m ca n b e s o lv e d , i f n u m b e r a g reem en t is con sid ered t o b e a p r e c o n d itio n fo r c o o r d in a tin g n ou n s. F or in sta n ce , h \u00f6g \u00f6verb yggn a d i f \u00f6 r e lle r a k te r p \u00e5 m ed eltid a fa r ty g [nn k a s te ll/2 ] ' h ig h s u p e rstru ctu re a t th e p ro w o r a t th e stern o f m ed iev a l v esse ls' . T w o ca ses, finally, czm b e so lv e d i f p h ra se o lo g y is ta k e n in to a c c o u n t , e .g . a k te r u t ' a s te r n ' . In T y p e 6 a n d T y p e 8 definiten ess is th e d istin g u ish in g fea tu re . It re a d ily so lv e s o n e o f th e th ree ca ses, i.e. v\u00e5g (t ill n g t) s o m e j f \u00f6 l je r d en n a tu rliga banan ' r o a d ( t o s o m e th in g ) w h ich d o e s n 't fo llo w th e n a tu ra l p a t h ' [nn b a k v \u00e4 g ]. A s reg a rd s defin iten ess in th e h ea d w ord o f p r e p o s itio n a l p h ra ses (t h e rem a in in g tw o ca s e s ), th e in d iv id u a l p re p o sitio n s m a k e th e ir o w n d e m a n d s, a n d th e situ a tio n is m o re c o m p lic a te d . F in ally, in T y p e 7, th e h o m o g r a p h c o m p o n e n ts d ifier w ith reg a rd t o ca se. It solv es the h o m o g r a p h y , i f th e g ra m m a r d o e s n ' t a d m it e llip tic N P h eads. avslu ta 'fin is h /c o n c lu d e ' 2 avslu tas ' fin is h (p a s s )/c o n c lu d e (p a s s )' 2 avlade ' b e g e t t e d /la y e d a s id e b le 10 w e p resen t th re e ty p e s o f v e rb h o m o g ra p h ie s . T h e first tw o have n o d istin g u ish in g fo r m a l fe a tu re , a n d th u s c a n n o t b e s o lv e d b y s y n ta c tic m ean s. In th e th ird ty p e , h ow e v e r, th e re is a c o n te x t in w h ich th e y d iffer, i.e. in past p a r tic ip le c o n s tr u c tio n s . In o u r m a te ria l, th is d is tin c tio n d o e s n ' t e m erg e, and th e h o m o g r a p h y c^lnnot b e s o lv e d sy n ta ctica lly .",
                "type_str": "figure"
            },
            "FIGREF16": {
                "uris": null,
                "num": null,
                "text": "d (c f. T a b le 1) c o m b in a tio n s o f th e h o m o g r a p h y ty p e s p resen te d a b o v e (c f. T a b le 5 ) w ere in v o lv e d . T h e a c tu a l c o m b in a tio n s a re p resen ted in T a b le 11, a n d T a b le 12.",
                "type_str": "figure"
            },
            "FIGREF17": {
                "uris": null,
                "num": null,
                "text": "ie v e d in a p ilo t s tu d y o f th e d e fin ition v o c a b u la r y o f S v en sk ordhok. It is p a rt o f o u r o n -g o in g w o rk o n th e g en e ra tion o f a m a c h in e -tr a c ta b le d ic tio n a r y fr o m th e le x ica l d a ta b a s e fr o m w h ich th e d ic tio n a r y w as d ra w n , in s p e cific, o f m a k in g its d e fin itio n s e x p lo ita b le t o th e L P S parser. P a rsin g th em is a ste p to w a rd s th is g o a l. T h e p re sen t stu d y , a im in g in p a rticu la r a t th e a u to m a t ic le m m a tis a tio n o f th e d e fin itio n v o c a b u la r y , has in clu d e d a n a u to m a tic m o r p h o lo g ic a l (in fle c tio n a l) a n a ly sis o f a su b set (2 ,5 0 0 g ra p h ic w o rd s ) o f it, an d a, la rg e ly m a n u a l, a n a ly sis o f th e resu lts. T w o m a jo r issues w ere tre a te d , i.e. th e c o v e r a g e o f th e d ic t io n a r y in re la tio n t o th e d e fin itio n v oa cu la ry , a n d th e fea sib ility o f h o m o g r a p h r e s o lu tio n b y s y n ta c tic a n a ly sis in th e lo c a l (d e fin itio n ) co n te x ts. T h e m o r p h o lo g ic a l a n a lysis w as ca rrie d o u t b y m ea n s o f th e L P S m o r p h o lo g ic a l a n a ly ser, d isp o sin g o f a ste m d ic t io n a r y c o v e r in g th e h ea d w o r d s o f th e d ictio n a ry . A s a resu lt o f this a n a ly sis, w e fo u n d th a t, ro u g h ly , 23 p e rce n t o f th e d efin ition w o rd s w ere n o t e x p lic itly d efin ed . T h e le x ic a l g a p s w e re m a in ly d u e to th e u se o f p r o p e r n ou n s (6 % ), a n d d e riv e d w o rd s (9 3 % ). A s reg a rd s th e m issin g p ro p e r n o u n s, th e y w ill b e re g istered in o u r m a ch in e d ictio n a ry . T h e d o m in a tin g ty p e o f th e d e riv e d w o rd s are th e c o m p o u n d s (7 9 % ), w h ile th e d eriv a tiv e s (in ch p a rticle v e rb s a n d p refix ed v e rb s ) a c c o u n t fo r 14 p e rce n t o f th e g a p s. T h e tra n s p a re n cy cla im w ith regard t o th e im p lic itly d e fin ed w o r d s Ctinnot b e m a in ta in e d w ith o u t ce rta in m o d ific a tio n s o f th e d e fin itio n v o ca b u la ry . T w o m easu res w ill b e taken in th e g e n e ra tio n o f th e m a ch in e d ic tio n a r y . F ir s t, d e riv e d w o rd s in c lu d in g h o m o g r a p h ic o r p o ly s e m o u s co n stitu e n ts w ill b e e x ch a n g e d b y e x p lic itly d efin e d w ord s, o r b y d e riv e d w o rd s o f u n a m b ig u o u s elem en ts. S e co n d ly , th e u se o f d e v e r bal n ou n s b a sed o n im p lic itly d efin ed p a r tic le v e rb s o r p re fix ed v e rb s , im p ly in g a sy ste m a tic a m b ig u ity , w ill b e a v o id e d , as w ill th e u se o f p re fix e d p a st p a rticip le s, in h eren tly a m b ig u o u s as th e y are. In g en era l, d e riv a tio n s o f m o r e th a n o n e ste p w ill b e a v o id e d . In o th e r w o rd s, w e req u ire th a t th e to p le v e l c o n s titu e n ts o f a c o m p o u n d o r a d e riv a tiv e are th em selv es e x p lic itly d efin ed . T h e s e m o d ific a tio n s , w h en a p p lie d to w h o le d efin ition v o c a b u la r y , w ill b rin g it o n e ste p clo s e r t o a d efin in g v oca b u la ry .",
                "type_str": "figure"
            },
            "TABREF0": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF1": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF2": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF3": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF4": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF5": {
                "content": "<table><tr><td>: Types o f N 1 / N 2 hom ographies.</td></tr></table>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF6": {
                "content": "<table><tr><td>T y p e</td><td>E xam ple</td><td>N u m ber o f</td></tr><tr><td/><td/><td>m em bers</td></tr><tr><td>N l/N 2 /A b /P p</td><td colspan=\"2\">bak 'behind/baking/at the beick/behind' 1</td></tr><tr><td>N 1 /N 2 /A /V</td><td>bar 'bar/bar/bare/carried'</td><td>1</td></tr><tr><td>T oted</td><td/><td/></tr></table>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            },
            "TABREF7": {
                "content": "<table/>",
                "num": null,
                "text": "",
                "html": null,
                "type_str": "table"
            }
        }
    }
}