From mcginnis at ucalgary.ca Mon Oct 1 15:14:48 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 09:14:48 -0600 Subject: Amanda Owen: Application of DM to error analyses Message-ID: I have been reading several things related to Distributed Morphology (specifically the DM FAQ, and a couple of papers suggested by my professor) as I consider the application of DM to a series of error analyses I have been conducting for data from Hebrew speaking children. I am tempted to draw on DM to support some of the hypotheses I am presenting, however I am relatively naive about the application of a morphological theory to what is in part a syntactic phenomenon. If these questions are too general or if there are references that answer these questions directly that I have missed, I apologize. Specifically, I am curious about how DM handles the presence of default (?non-distinct) features. I understand, I think, that some forms are less marked than others and that these are more likely to be chosen. Thus if a morphological inflection is required and the features are not clearly specified, then the least-marked form that most matches without conflicting would be chosen. I think this is underspecification, but I am unclear about how it would actually be sorted out. If features have been argued to reside on multiple XPs (Ritter, 1995... in Hebrew, Number and Gender reside below TP while Person and Definiteness are located above TP; all of these elements are argued to be overtly or covertly realized on the verb through infixes, prefixes, and suffixes), however the realization of those features phonetically is in one composite overt form, is this contradictory within DM? Can one element lower in the tree be in error (use an unmarked form) while features higher in the tree are completely specified? Ideally, I would like to think that the process of probe, agree, copy and delete argued for in Derivation by Phase (chomsky) would apply first, with DM then completing the phonological representation, but I am not sure if thinking in terms of orders is accurate nor am I sure that the two methods of addressing features are not contradictory. Finally, within DM how would one identify the default features for a particular language? i.e. first acquired? least phonologically marked? least errors? what is the criterion generally used? Thank you very much for any assistance that might be available in dealing with these questions. Amanda Owen ajowen at purdue.edu Doctoral Student Dept of Audiology and Speech Sciences Purdue University, Indiana USA From mcginnis at ucalgary.ca Mon Oct 1 18:18:00 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 12:18:00 -0600 Subject: Carson Schutze: Application of DM to error analyses (reply to Amanda Owen) Message-ID: Hi Amanda, Might I humbly refer you to my dissertation? It discusses the nature of defaults in DM in the context of acquisition, as well as adult syntax, concentrating on case and agreement. There is also some discussion on the question I think you're alluding to near the end of your message, namely how syntactic feature checking does or doesn't interact with morphological defaults. I've changed my thinking on that a bit, am now leaning more towards the Marantzian view that says that the whole case & agreement system really belongs in the spellout component, not in the narrow syntax, but in the diss. you'll see an attempt to argue the opposite. As Alec said to me recently, this issue is "very tricky" :-) You might also be interested in a paper about to appear in the journal Syntax, which expands on my dissertation discussion about default case. Here are the references: Schütze, Carson T. (1997) INFL in child and adult language: Agreement, case and licensing. Doctoral dissertation, MIT. Distributed by MIT Working Papers in Linguistics. --- (2001) On the nature of default case. In press, Syntax 4(3). I would be happy to discuss these issues with you in more detail. Carson -- Carson T. Schutze Department of Linguistics, UCLA Email: cschutze at ucla.edu Box 951543, Los Angeles CA 90095-1543 U.S.A. Office: Campbell Hall 2224B Deliveries/Courier: 3125 Campbell Hall Campus Mail Code: 154302 Web: www.linguistics.ucla.edu/people/cschutze Phone: (310)995-9887 Fax: (310)206-8595 -------------------------------------------------------------------------- "Words and music. I love them. And especially what happens when you put them together into songs. And sing them, in a large building, in a central part of town, as part of a play, with a lot of people listening, who have all paid a great deal to get in." --Ed Kleban, lyricist of 'A Chorus Line' From mcginnis at ucalgary.ca Mon Oct 1 19:11:23 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 13:11:23 -0600 Subject: Rolf Noyer: Application of DM to error analyses (reply to Amanda Owen) Message-ID: Amanda, In addition to what Carson said, you might want to take a look at some of the following papers: Halle, Morris. 1992. 'Latvian Declension.' Morphology Yearbook 1991, ed. Geert Booij and Jaap van der Marle. Kluwer, Dordrecht, 33-47. Halle, Morris. 1995. 'The Russian Declension.' In Perspectives in Phonology, ed. Jennifer Cole and Charles Kisseberth. CSLI, Stanford, pp. 321-353. Halle, Morris. 1997. 'Distributed morphology: Impoverishment and fission.' In MITWPL 30: Papers at the Interface, ed. Benjamin Bruening, Yoonjung Kang and Martha McGinnis. MITWPL, Cambridge, 425-449. Noyer, Rolf. 1998. 'Impoverishment theory and morphosyntactic markedness.' In Morphology and its relation to phonology and syntax, ed. Steve Lapointe, Diane K. Brentari, and Patrick Farrell. CSLI, Stanford, 264-285. Harris, James. 1997a. 'Why n'ho is pronounced [li] in Barceloni Catalan.' In MITWPL 30: Papers at the Interface, ed. Benjamin Bruening, Yoonjung Kang and Martha McGinnis. MITWPL, Cambridge, 451-479. ...for some discussion of default rules in DM. There are several ways in which forms which are "unmarked" can show up. They are: (1) The form is a default spell-out within the system of spell-out rules. Therefore, in the absence of any features, this form is inserted. (2) The morphological features which trigger the spell-out of the form are themselves defaults/"unmarked". So for, say, Spanish, according to Jim Harris, a default rules supplies to unmarked masculine nouns the declensional class I (or the equivalent) which ends up meaning that they receive the word marker /-o/. Nouns may be masculine and not have the word marker /-o/ -- these must be specified as some other class, and this specification overrides the default class assignment rule. Likewise, nouns do not have to have their class feature supplied by the default rule: a noun such as /man-o/ 'hand' is feminine but specified as class I. (3) Impoverishment may operate in certain contexts making what would otherwise be a "marked" form not appear, and allowing instead an "unmarked" form. Again, the example for Spanish that Jim Harris proposed involves the sequence of 3rd person indirect object clitics followed by 3rd person object clitics. For the indirect object clitic the expected form is replaced in this context by an "unmarked" form /s-(e)/. On the matter of combining various functional projections such as Num, Agr, and even Gender (not all of which are necessarily functional projections -- and on such an issue DM is silent since the precise inventory of projections is not crucial to the tenets of the theory), if one assumes that these are separate in syntax, then each should, by default, be a morpheme subject to spell-out as an individual affix. So to anyone wedded to such an expanded INFL approach the first thing to look for would be evidence from the morphology that INFL does indeed have a tripartite (bipartite) structure. If this proves fruitless there is the possibility of Fusing functional heads as single morphemes, although on this point there has been little overall agreement, I think. This was the approach I used in my thesis (fusing T and AGR) prior to spell-out in Arabic. There you can also find discussion of the Hebrew prefix conjugation; for an alternative, see Halle 1997, for example. Rolf Noyer From mcginnis at ucalgary.ca Mon Oct 1 15:14:48 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 09:14:48 -0600 Subject: Amanda Owen: Application of DM to error analyses Message-ID: I have been reading several things related to Distributed Morphology (specifically the DM FAQ, and a couple of papers suggested by my professor) as I consider the application of DM to a series of error analyses I have been conducting for data from Hebrew speaking children. I am tempted to draw on DM to support some of the hypotheses I am presenting, however I am relatively naive about the application of a morphological theory to what is in part a syntactic phenomenon. If these questions are too general or if there are references that answer these questions directly that I have missed, I apologize. Specifically, I am curious about how DM handles the presence of default (?non-distinct) features. I understand, I think, that some forms are less marked than others and that these are more likely to be chosen. Thus if a morphological inflection is required and the features are not clearly specified, then the least-marked form that most matches without conflicting would be chosen. I think this is underspecification, but I am unclear about how it would actually be sorted out. If features have been argued to reside on multiple XPs (Ritter, 1995... in Hebrew, Number and Gender reside below TP while Person and Definiteness are located above TP; all of these elements are argued to be overtly or covertly realized on the verb through infixes, prefixes, and suffixes), however the realization of those features phonetically is in one composite overt form, is this contradictory within DM? Can one element lower in the tree be in error (use an unmarked form) while features higher in the tree are completely specified? Ideally, I would like to think that the process of probe, agree, copy and delete argued for in Derivation by Phase (chomsky) would apply first, with DM then completing the phonological representation, but I am not sure if thinking in terms of orders is accurate nor am I sure that the two methods of addressing features are not contradictory. Finally, within DM how would one identify the default features for a particular language? i.e. first acquired? least phonologically marked? least errors? what is the criterion generally used? Thank you very much for any assistance that might be available in dealing with these questions. Amanda Owen ajowen at purdue.edu Doctoral Student Dept of Audiology and Speech Sciences Purdue University, Indiana USA From mcginnis at ucalgary.ca Mon Oct 1 18:18:00 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 12:18:00 -0600 Subject: Carson Schutze: Application of DM to error analyses (reply to Amanda Owen) Message-ID: Hi Amanda, Might I humbly refer you to my dissertation? It discusses the nature of defaults in DM in the context of acquisition, as well as adult syntax, concentrating on case and agreement. There is also some discussion on the question I think you're alluding to near the end of your message, namely how syntactic feature checking does or doesn't interact with morphological defaults. I've changed my thinking on that a bit, am now leaning more towards the Marantzian view that says that the whole case & agreement system really belongs in the spellout component, not in the narrow syntax, but in the diss. you'll see an attempt to argue the opposite. As Alec said to me recently, this issue is "very tricky" :-) You might also be interested in a paper about to appear in the journal Syntax, which expands on my dissertation discussion about default case. Here are the references: Sch?tze, Carson T. (1997) INFL in child and adult language: Agreement, case and licensing. Doctoral dissertation, MIT. Distributed by MIT Working Papers in Linguistics. --- (2001) On the nature of default case. In press, Syntax 4(3). I would be happy to discuss these issues with you in more detail. Carson -- Carson T. Schutze Department of Linguistics, UCLA Email: cschutze at ucla.edu Box 951543, Los Angeles CA 90095-1543 U.S.A. Office: Campbell Hall 2224B Deliveries/Courier: 3125 Campbell Hall Campus Mail Code: 154302 Web: www.linguistics.ucla.edu/people/cschutze Phone: (310)995-9887 Fax: (310)206-8595 -------------------------------------------------------------------------- "Words and music. I love them. And especially what happens when you put them together into songs. And sing them, in a large building, in a central part of town, as part of a play, with a lot of people listening, who have all paid a great deal to get in." --Ed Kleban, lyricist of 'A Chorus Line' From mcginnis at ucalgary.ca Mon Oct 1 19:11:23 2001 From: mcginnis at ucalgary.ca (Martha McGinnis) Date: Mon, 1 Oct 2001 13:11:23 -0600 Subject: Rolf Noyer: Application of DM to error analyses (reply to Amanda Owen) Message-ID: Amanda, In addition to what Carson said, you might want to take a look at some of the following papers: Halle, Morris. 1992. 'Latvian Declension.' Morphology Yearbook 1991, ed. Geert Booij and Jaap van der Marle. Kluwer, Dordrecht, 33-47. Halle, Morris. 1995. 'The Russian Declension.' In Perspectives in Phonology, ed. Jennifer Cole and Charles Kisseberth. CSLI, Stanford, pp. 321-353. Halle, Morris. 1997. 'Distributed morphology: Impoverishment and fission.' In MITWPL 30: Papers at the Interface, ed. Benjamin Bruening, Yoonjung Kang and Martha McGinnis. MITWPL, Cambridge, 425-449. Noyer, Rolf. 1998. 'Impoverishment theory and morphosyntactic markedness.' In Morphology and its relation to phonology and syntax, ed. Steve Lapointe, Diane K. Brentari, and Patrick Farrell. CSLI, Stanford, 264-285. Harris, James. 1997a. 'Why n'ho is pronounced [li] in Barceloni Catalan.' In MITWPL 30: Papers at the Interface, ed. Benjamin Bruening, Yoonjung Kang and Martha McGinnis. MITWPL, Cambridge, 451-479. ...for some discussion of default rules in DM. There are several ways in which forms which are "unmarked" can show up. They are: (1) The form is a default spell-out within the system of spell-out rules. Therefore, in the absence of any features, this form is inserted. (2) The morphological features which trigger the spell-out of the form are themselves defaults/"unmarked". So for, say, Spanish, according to Jim Harris, a default rules supplies to unmarked masculine nouns the declensional class I (or the equivalent) which ends up meaning that they receive the word marker /-o/. Nouns may be masculine and not have the word marker /-o/ -- these must be specified as some other class, and this specification overrides the default class assignment rule. Likewise, nouns do not have to have their class feature supplied by the default rule: a noun such as /man-o/ 'hand' is feminine but specified as class I. (3) Impoverishment may operate in certain contexts making what would otherwise be a "marked" form not appear, and allowing instead an "unmarked" form. Again, the example for Spanish that Jim Harris proposed involves the sequence of 3rd person indirect object clitics followed by 3rd person object clitics. For the indirect object clitic the expected form is replaced in this context by an "unmarked" form /s-(e)/. On the matter of combining various functional projections such as Num, Agr, and even Gender (not all of which are necessarily functional projections -- and on such an issue DM is silent since the precise inventory of projections is not crucial to the tenets of the theory), if one assumes that these are separate in syntax, then each should, by default, be a morpheme subject to spell-out as an individual affix. So to anyone wedded to such an expanded INFL approach the first thing to look for would be evidence from the morphology that INFL does indeed have a tripartite (bipartite) structure. If this proves fruitless there is the possibility of Fusing functional heads as single morphemes, although on this point there has been little overall agreement, I think. This was the approach I used in my thesis (fusing T and AGR) prior to spell-out in Arabic. There you can also find discussion of the Hebrew prefix conjugation; for an alternative, see Halle 1997, for example. Rolf Noyer