============================================= Sentential Negation in Malayalam ============================================= Malayalam has 2 'being' verbs: aanax and alla. Each of these has a negative form. aanax -> unatax alla -> illa Sentences which use these verbs, replace them with the negative form to create sentential negation. raaman dooktaar aanax Raman doctor be-PRES Raman is a doctor raaman dooktaar alla Raman doctor be-PRES-NEG Raman is not a doctor kutatai viatail unatax child house-LOC be-PRES the child is at home kutatai viatail illa child house-LOC be-PRES-NEG the child is not at home Sentences with verbs other than these 2, append illa or alla to the end of the verb to create a negative sentence. avan patahiccu -> he studied avan patahicilla -> he didn't study In some cases, either alla or illa can be used with the same verb. I'm not sure exactly what the nuance is between the 2 verbs, but aanax implies 'is' while unatax is more like 'has'. avannx pani aanax / alla -> he is/isn't feverish avannx pani unatax / illa -> he has/doesn't have a fever ============================================= Implementation of Sentential Negation (Description of how it works) ============================================= The lexical rule from the creation script works for my sentential negation. neg-infl-lex-rule := cont-change-only-lex-rule & inflecting-lex-rule & [ C-CONT [ HOOK [ XARG #xarg, LTOP #ltop, INDEX #ind ], RELS , HCONS ], SYNSEM.LKEYS #lkeys, DTR lex-item & [ SYNSEM [ LKEYS #lkeys, LOCAL [ CONT.HOOK [ XARG #xarg, LTOP #larg ], CAT.HEAD verb ] ] ] ] . HOOK has a feature XARG which is the pointer to the oject being change to the negative form. That corresponds to the daughter (dowstairs verb). HOOK.LTOP is the handle of the result of the lexical rule and becomes the handle which labels the RELS predication. HOOK.INDEX is the local sign, and becomes the first argument in the RELS predication. RELS includes the label of the relation, neg_r_rel, as well as a list of arguments to the relation. ARG0 is the mother, and ARG1 is the handle of the higher scope element of the qeq list in HCONS. The higher scope element in HCONS is related to the lower scope element, which is the handle of the daughter in the rule. So the input is a verb, and this is then related to the output which is an inflected verb, but semantically relates the verb to its negative form. A semantic negative is added to the original verb's semantic meaning. I have created 2 instances of this inflecton, since the negation can take place with either appending illa or alla to a verb. Since I don't fully understand the difference between the 2 'be' verbs and their negative form, I don't have the means to limit their use in generation to only the appropriate case. So these rules are mainly to get the correct parsing. They won't be helpful in generation. neg1-infl-lr := %suffix (* alla) neg-infl-lex-rule. neg2-infl-lr := %suffix (* illa) neg-infl-lex-rule. ============================================= Current Coverage: Does it get the right strings and only the right strings. Does it get the right meaning. Are there any problems, and what could be done? ============================================= This rule, and the inflections are parsing the sentences, and getting the meaning. After looking at the MRS, I think this is what's happening. The MRS is showing some elements out of both CONT and HOOK? The LTOP handle is the handle of the sentence. The first node is the negation (ARG0) of the sentence (ARG1) which has handle 15 (in the MRS diagram I have) Handle 15 is related to handle 13 though a qeq relationship, where h13 is the verb which heads the sentence. h13 then has 3 ARGS: ARG0: an inflection ARG1: the head daughter (x4-> h6 -> qeq -> h3: it) ARG2: complement (x9 -> h11 -> qeq -> h8: me) This seems very straightforawrd, with no problems. The generation does not work. I get a result that no strings were generated, and that the edge limit was exhausted. ============================================= Describe 'can' in Malayalam ============================================= Malayalam includes a modal auxiliary which is in the form of the suxxif -aam. Nair's book "Auxiliary Verbs in Malayalam" says that this is a contracted form of the future tense form of the copula verb 'to be'. The suffic -aam can be attached to the verb in 2 different syntactic structures: I. subj-DAT + verb-aam II. subj-NOM + verb-aam In form I, the suffix acts as an auxiliary verb and denotes permission, capability, possibility, inclination or suggestion. In form II, it acts as a main verb and can mean probibility, certainty, or willingness. There are also 4 verbs which can denote capability by themselves: kaliyum parrum okkum saadhikkum Thus a sentence can be structured as ninnalakkx codyannala codikkaam you-DAT questions ask-can or ninnalakkx codyannala codikkan kaliyum you-DAT questions ask can For my test cases, I have used the suffix since it seems to be more common. ============================================= Describe the implementation ============================================= I used the suggested lexical rule from the homework for the auxiliary suffix. It contains many of the same contraints as the negation rule above. The verb is identified as the top handle. As input, it takes an uninflected verb, since the suffix is added to the infinitive form of the verb. HOOK's XARG feature is a pointer to the object being controlled. LTOP identifies this element as the top handle, and the INDEX is identified with ARG0 of the relationship. RELS has the 2 arguments to the relation: ARG0 which is the output of the rule, and ARG1, which is the upstairs element of the qeq relation. HCONS identifies the qeq relation, and the LARG with the daughter of the rule. ; Auxiliary Verb infl-add-ccont-ltow-rule := inflecting-lex-rule & same-non-local-lex-rule & same-cat-lex-rule & same-ctxt-lex-rule & same-agr-lex-rule & [ INFLECTED +, C-CONT [ HOOK [ XARG #xarg, LTOP #ltop, INDEX #ind ], RELS , HCONS ], DTR [ SYNSEM.LOCAL [CONT.HOOK [XARG #xarg, LTOP #larg ], CAT.HEAD verb ], INFLECTED - ] ]. The inflection adds -aam to words that end in a consonant, or if the word ends in a, the final a is dropped before the suffix -aam is added. aux-infl-verb := %suffix (* aam) (!xa !xaam) infl-add-ccont-ltow-rule. ============================================= Current Coverage: Does it get the right strings and only the right strings. Does it get the right meaning. Are there any problems, and what could be done? ============================================= This rule, is parsing the sentences, and getting the meaning. Can's ARG1 is qeq to the handle of eat as expected. Eat in turn has 2 relations: x4 being the index of the personal pronoun, and x9 the index of glass. This seems very straightforawrd, with no problems. The generation does not work. I get a result that no strings were generated, and that the edge limit was exhausted. ============================================= What else would I like to fix? ============================================= Malayalam has an interesting nuance where the object is denoted by case in the sentence. The object's case however, is affected by its type. If the object is HUMAN or ANIMAL (I think basically if it's something that's sentient) then its case is ACC. Otherwise the object's case is NOM. I'd like to revise my rules to account for this.