From anghelutar at YAHOO.COM Thu Aug 19 18:49:42 2004 From: anghelutar at YAHOO.COM (Roxana Angheluta (by way of Martha McGinnis)) Date: Thu, 19 Aug 2004 12:49:42 -0600 Subject: tools and resources Message-ID: Deal all, My name is Roxana Angheluta, I graduated Computer Science and I'm currently researcher in the domain of Information Retrieval at the Katholieke Universiteit Leuven, Belgium. Seeing the domain of nlp from the perspective of an engineer, not a linguist, I encounter many problems because of the lack of tools/resources freely available for the community. I have the feeling that in practice we are far behind what has been achieved from a theoretical perspective. At the moment, I'm interested in resources/tools which would allow me to identify semantically related words from the same morphological family, e.g.: label/V, label/N, labeled/Adj, labelled/Adj. I'm not interested only in the words with the same stem, but rather to a larger class, which includes words like measure/N and measurement/N. Can anyone point to me such resources? I've already tried 2 of them: WordNet and CELEX database. The problem with WordNet is that is doesn't have many pointers accross different syntactical categories, and alowing a longer path between related words introduces a lot of noise, making it impractical for this specific problem. The other resource I've seen, CELEX database, does the job, but is limited in the number of lexemes/spellings it contains (e.g. it does not contain the spelling "labelled"). I appologize if this is not the right place for this topic. Please reply to my personal email (anghelutar @ yahoo.com), since I'm not subscribed to this list. Thanks, roxana From DM-LIST at LISTSERV.LINGUISTLIST.ORG Thu Aug 26 10:06:32 2004 From: DM-LIST at LISTSERV.LINGUISTLIST.ORG (Ora Matushansky) Date: Thu, 26 Aug 2004 12:06:32 +0200 Subject: query Message-ID: Dear everyone, Does anyone know of morphological complexity blocking incorporation, in particular in sign languages? Many thanks in advance, O Ora Matushansky CNRS - UMR 7023 (Paris 8) homepage: http://mapage.noos.fr/matushan/ ---------- Please note the change in the email address! SVP notez que mon adresse mél a changé! From R.Pfau at UVA.NL Mon Aug 30 12:14:04 2004 From: R.Pfau at UVA.NL (Pfau, R.) Date: Mon, 30 Aug 2004 14:14:04 +0200 Subject: query incorporation Message-ID: Dear Ora, I would like to say a few words concerning the sign language part of your question. In the sign language literature, people make abundant use of the term "incorporation". This is somewhat misleading, since in most of the cases the term is not used in its technical sense. 1. One example is "negative incorporation". Some verbs (above all modal verbs) may change there movement properties (and/of other phonological properties) when negated. This, however, is probably due to cliticization or suppletion, not to incorporation. In any case, it is not relevant for your question, since the negative elements are not morphologically complex. 2. Another case is "object & subject incorporation". This is sometimes used in the context of classifying verb which classify one of there arguments by means of a handshape change (e.g. GIVE). Here the use of "incorporation" is certainly wrong, since the argument is never incorporated. It is always present verb-externally, the handshape only reflects a characteristic of the subject/object, as in (i): (i) MAN WOMAN FLOWER GIVE-CL(long/thin) where FLOWER always has to be signed seperately. Dropping FLOWER leads to ungrammaticality (the same is true for subject classifiers in intransitive sentences). Therefore, some people have argued that the classifier is a noun class or agreement marker (Zwitserlood 2003, Glueck & Pfau 1998). There may be one exception, however. Meir (2001) argues for Israeli Sign Language that some instances of classification, viz. instrumental classifiers, are in fact incorporation. The example she gives is I BABY INDEXa SPOON-FEEDa where addition of the sign SPOON would lead to ungrammaticality. Now, I don't think I agree with her proposal, but for the sake of the argument, let's assume that she is right. Let's further assume - following Supalla (1986) and Zwitserlood (2003), among others - that classifier handshapes may be morphologically complex in that every finger (or even joint) may have a morphemic function. One could then in principle argue that a morphologically complex element has been incorporated, although Meir does not give examples of that sort. (In fact, she does not even give a FORK-FEED example; maybe she's afraid one might hurt the baby.) In that sense, morphological complexity does not block incorporation. 3. The third area where the term "incorporation" is commonly used is "numeral incorporation" and here, I think, the term is appropriately used. In sign languages, numerals frequently incorporate into nouns that refer to time, weight, money etc but also into the signs for HUNDRED and THOUSAND. Incorporation, however, is clearly constrained. Some of these constraints are predictable, others are idiosyncratic. Moreover, sign languages differ from each other with respect to what is allowed and what is blocked. In Sign Language of the Netherlands (NGT) (and many other SLs) the following alternatives exist. (i) I WAIT TWO WEEK (ii) I WAIT TWO-WEEK where in the second example only the location and the movement of WEEK is retained but it is signed with the 2-handshape. Since NGT uses a one-handed counting system, numbers up to nine can be incorporated. Numbers from 11 to 19 in NGT are not morphologically complex, only a movement is added. And it is this movement that blocks incorporation of these numbers (for Chinese SL it has been argued that higher numbers can incorporate). Movement also changes for the signs for 20, 22, 30, 33, 40, 44 etc. Only numbers like 21, 23 etc are morphologically complex in that they are compounds of ONE-TWENTY, THREE-TWENTY etc. And obviuosly, these numbers cannot incorporate either. Therefore, here one could argue that morphological complexity blocks incorporation. But given that movement also blocks incorporation, the blocking might of course also be due to phonological factors. A final note: due to idiosyncratic gaps in the system, it has been argued for ASL (Liddell 1997) that all of these complex signs are lexicalized anyway. In sum, it looks like we don't find clear cases in sign langauges (at least, those that I am familiar with) where morphological complexity blocks incorporation. Ora, I hope this is helpful. Best, Roland. REFERENCES: Glueck, S. & R. Pfau (1998), On classifying classification as a class of inflection in German Sign Language. In: Cambier-Langeveld, T., A. Lipták & M. Redford (eds.), Proceedings of ConSole 6. Leiden: SOLE, 59-74 Liddell, S.K. (1997), Numeral incorporating roots & non-incorporating prefixes in American Sign Language. Sign Language Studies 92, 201-225. Meir, I. (2001), Verb classifiers as noun incorporation in Israeli Sign Language. In: Booij, G. & J. van Marle (eds.), Yearbook of Morphology 1999. Dordrecht: Kluwer, 299-319. Supalla, T. (1986), The classifier system in American Sign Language. In: Craig, C. (ed.), Noun classes and categorization. Amsterdam: Benjamins, 181-214. Zwitserlood, I. (2003), Classifiying hand configurations in Nederlandse Gebarentaal. Utrecht: LOT. ********************************* Dr. Roland Pfau Assistant Professor Dept. of General Linguistics University of Amsterdam Spuistraat 210 1012 VT Amsterdam The Netherlands tel. 0031-(0)20-5253022 r.pfau at uva.nl ********************************* -----Original Message----- From: The Distributed Morphology List [mailto:DM-LIST at LISTSERV.LINGUISTLIST.ORG]On Behalf Of Ora Matushansky Sent: donderdag 26 augustus 2004 12:07 To: DM-LIST at LISTSERV.LINGUISTLIST.ORG Subject: query Dear everyone, Does anyone know of morphological complexity blocking incorporation, in particular in sign languages? Many thanks in advance, O Ora Matushansky From anghelutar at YAHOO.COM Thu Aug 19 18:49:42 2004 From: anghelutar at YAHOO.COM (Roxana Angheluta (by way of Martha McGinnis)) Date: Thu, 19 Aug 2004 12:49:42 -0600 Subject: tools and resources Message-ID: Deal all, My name is Roxana Angheluta, I graduated Computer Science and I'm currently researcher in the domain of Information Retrieval at the Katholieke Universiteit Leuven, Belgium. Seeing the domain of nlp from the perspective of an engineer, not a linguist, I encounter many problems because of the lack of tools/resources freely available for the community. I have the feeling that in practice we are far behind what has been achieved from a theoretical perspective. At the moment, I'm interested in resources/tools which would allow me to identify semantically related words from the same morphological family, e.g.: label/V, label/N, labeled/Adj, labelled/Adj. I'm not interested only in the words with the same stem, but rather to a larger class, which includes words like measure/N and measurement/N. Can anyone point to me such resources? I've already tried 2 of them: WordNet and CELEX database. The problem with WordNet is that is doesn't have many pointers accross different syntactical categories, and alowing a longer path between related words introduces a lot of noise, making it impractical for this specific problem. The other resource I've seen, CELEX database, does the job, but is limited in the number of lexemes/spellings it contains (e.g. it does not contain the spelling "labelled"). I appologize if this is not the right place for this topic. Please reply to my personal email (anghelutar @ yahoo.com), since I'm not subscribed to this list. Thanks, roxana From DM-LIST at LISTSERV.LINGUISTLIST.ORG Thu Aug 26 10:06:32 2004 From: DM-LIST at LISTSERV.LINGUISTLIST.ORG (Ora Matushansky) Date: Thu, 26 Aug 2004 12:06:32 +0200 Subject: query Message-ID: Dear everyone, Does anyone know of morphological complexity blocking incorporation, in particular in sign languages? Many thanks in advance, O Ora Matushansky CNRS - UMR 7023 (Paris 8) homepage: http://mapage.noos.fr/matushan/ ---------- Please note the change in the email address! SVP notez que mon adresse m?l a chang?! From R.Pfau at UVA.NL Mon Aug 30 12:14:04 2004 From: R.Pfau at UVA.NL (Pfau, R.) Date: Mon, 30 Aug 2004 14:14:04 +0200 Subject: query incorporation Message-ID: Dear Ora, I would like to say a few words concerning the sign language part of your question. In the sign language literature, people make abundant use of the term "incorporation". This is somewhat misleading, since in most of the cases the term is not used in its technical sense. 1. One example is "negative incorporation". Some verbs (above all modal verbs) may change there movement properties (and/of other phonological properties) when negated. This, however, is probably due to cliticization or suppletion, not to incorporation. In any case, it is not relevant for your question, since the negative elements are not morphologically complex. 2. Another case is "object & subject incorporation". This is sometimes used in the context of classifying verb which classify one of there arguments by means of a handshape change (e.g. GIVE). Here the use of "incorporation" is certainly wrong, since the argument is never incorporated. It is always present verb-externally, the handshape only reflects a characteristic of the subject/object, as in (i): (i) MAN WOMAN FLOWER GIVE-CL(long/thin) where FLOWER always has to be signed seperately. Dropping FLOWER leads to ungrammaticality (the same is true for subject classifiers in intransitive sentences). Therefore, some people have argued that the classifier is a noun class or agreement marker (Zwitserlood 2003, Glueck & Pfau 1998). There may be one exception, however. Meir (2001) argues for Israeli Sign Language that some instances of classification, viz. instrumental classifiers, are in fact incorporation. The example she gives is I BABY INDEXa SPOON-FEEDa where addition of the sign SPOON would lead to ungrammaticality. Now, I don't think I agree with her proposal, but for the sake of the argument, let's assume that she is right. Let's further assume - following Supalla (1986) and Zwitserlood (2003), among others - that classifier handshapes may be morphologically complex in that every finger (or even joint) may have a morphemic function. One could then in principle argue that a morphologically complex element has been incorporated, although Meir does not give examples of that sort. (In fact, she does not even give a FORK-FEED example; maybe she's afraid one might hurt the baby.) In that sense, morphological complexity does not block incorporation. 3. The third area where the term "incorporation" is commonly used is "numeral incorporation" and here, I think, the term is appropriately used. In sign languages, numerals frequently incorporate into nouns that refer to time, weight, money etc but also into the signs for HUNDRED and THOUSAND. Incorporation, however, is clearly constrained. Some of these constraints are predictable, others are idiosyncratic. Moreover, sign languages differ from each other with respect to what is allowed and what is blocked. In Sign Language of the Netherlands (NGT) (and many other SLs) the following alternatives exist. (i) I WAIT TWO WEEK (ii) I WAIT TWO-WEEK where in the second example only the location and the movement of WEEK is retained but it is signed with the 2-handshape. Since NGT uses a one-handed counting system, numbers up to nine can be incorporated. Numbers from 11 to 19 in NGT are not morphologically complex, only a movement is added. And it is this movement that blocks incorporation of these numbers (for Chinese SL it has been argued that higher numbers can incorporate). Movement also changes for the signs for 20, 22, 30, 33, 40, 44 etc. Only numbers like 21, 23 etc are morphologically complex in that they are compounds of ONE-TWENTY, THREE-TWENTY etc. And obviuosly, these numbers cannot incorporate either. Therefore, here one could argue that morphological complexity blocks incorporation. But given that movement also blocks incorporation, the blocking might of course also be due to phonological factors. A final note: due to idiosyncratic gaps in the system, it has been argued for ASL (Liddell 1997) that all of these complex signs are lexicalized anyway. In sum, it looks like we don't find clear cases in sign langauges (at least, those that I am familiar with) where morphological complexity blocks incorporation. Ora, I hope this is helpful. Best, Roland. REFERENCES: Glueck, S. & R. Pfau (1998), On classifying classification as a class of inflection in German Sign Language. In: Cambier-Langeveld, T., A. Lipt?k & M. Redford (eds.), Proceedings of ConSole 6. Leiden: SOLE, 59-74 Liddell, S.K. (1997), Numeral incorporating roots & non-incorporating prefixes in American Sign Language. Sign Language Studies 92, 201-225. Meir, I. (2001), Verb classifiers as noun incorporation in Israeli Sign Language. In: Booij, G. & J. van Marle (eds.), Yearbook of Morphology 1999. Dordrecht: Kluwer, 299-319. Supalla, T. (1986), The classifier system in American Sign Language. In: Craig, C. (ed.), Noun classes and categorization. Amsterdam: Benjamins, 181-214. Zwitserlood, I. (2003), Classifiying hand configurations in Nederlandse Gebarentaal. Utrecht: LOT. ********************************* Dr. Roland Pfau Assistant Professor Dept. of General Linguistics University of Amsterdam Spuistraat 210 1012 VT Amsterdam The Netherlands tel. 0031-(0)20-5253022 r.pfau at uva.nl ********************************* -----Original Message----- From: The Distributed Morphology List [mailto:DM-LIST at LISTSERV.LINGUISTLIST.ORG]On Behalf Of Ora Matushansky Sent: donderdag 26 augustus 2004 12:07 To: DM-LIST at LISTSERV.LINGUISTLIST.ORG Subject: query Dear everyone, Does anyone know of morphological complexity blocking incorporation, in particular in sign languages? Many thanks in advance, O Ora Matushansky