{ "paper_id": "2020", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T04:33:37.257354Z" }, "title": "Part-of-Speech Annotation Challenges in Marathi", "authors": [ { "first": "Gajanan", "middle": [], "last": "Rane", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "" }, { "first": "Nilesh", "middle": [], "last": "Joshi", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "" }, { "first": "Geetanjali", "middle": [], "last": "Rane", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "" }, { "first": "Hanumant", "middle": [], "last": "Redkar", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "" }, { "first": "Malhar", "middle": [], "last": "Kulkarni", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "" }, { "first": "Pushpak", "middle": [], "last": "Bhattacharyya", "suffix": "", "affiliation": { "laboratory": "", "institution": "Indian Institute of Technology Bombay", "location": { "addrLine": "India {gkrane45, joshinilesh60, geetanjaleerane, hanumantredkar", "settlement": "Mumbai" } }, "email": "pushpakbh@gmail.com" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "Part of Speech (POS) annotation is a significant challenge in natural language processing. The paper discusses issues and challenges faced in the process of POS annotation of the Marathi data from four domains viz., tourism, health, entertainment and agriculture. During POS annotation, a lot of issues were encountered. Some of the major ones are discussed in detail in this paper. Also, the two approaches viz., the lexical (L approach) and the functional (F approach) of POS tagging have been discussed and presented with examples. Further, some ambiguous cases in POS annotation are presented in the paper.", "pdf_parse": { "paper_id": "2020", "_pdf_hash": "", "abstract": [ { "text": "Part of Speech (POS) annotation is a significant challenge in natural language processing. The paper discusses issues and challenges faced in the process of POS annotation of the Marathi data from four domains viz., tourism, health, entertainment and agriculture. During POS annotation, a lot of issues were encountered. Some of the major ones are discussed in detail in this paper. Also, the two approaches viz., the lexical (L approach) and the functional (F approach) of POS tagging have been discussed and presented with examples. Further, some ambiguous cases in POS annotation are presented in the paper.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "In any natural language, Part of Speech (POS) such as noun, pronoun, adjective, verb, adverb, demonstrative, etc., forms an integral building block of a sentence structure. POS tagging 1 is one of the major activities in Natural Language Processing (NLP). In corpus linguistics, POS tagging is the process of marking/annotating a word in a text/corpus which corresponds to a particular POS. The annotation is done based on its definition and its context i.e., its relationship with adjacent and related words in a phrase, sentence, or paragraph. A simplified form of this is commonly taught to school-age children, in the identification of words as nouns, verbs, adjectives, adverbs, etc. The term 'Part-of-Speech Tagging' is also known as POS tagging, POST, POS annotation, grammatical tagging or word-category disambiguation.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "In this paper, the challenges and issues in POS tagging with special reference to Marathi 2 have been presented. The Marathi language is one of the major languages of India. It belongs to the Indo-Aryan Language family with about 71,936,894 users 3 . It is predominantly spoken in the state of Maharashtra in Western India (Chaudhari et. al., 2017) . In recent years many research institutions and organizations are involved in developing the lexical resources for Marathi for NLP activities. Marathi Wordnet is one such lexical resource developed at IIT Bombay (Popale and Bhattacharyya, 2017) .", "cite_spans": [ { "start": 323, "end": 348, "text": "(Chaudhari et. al., 2017)", "ref_id": "BIBREF1" }, { "start": 562, "end": 594, "text": "(Popale and Bhattacharyya, 2017)", "ref_id": "BIBREF5" } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "The paper is organized as follows: Section 2 introduces POS annotation; section 3 provides information on Marathi annotated corpora; section 4 describes Marathi tag set; section 5 explains tagging approaches, section 6 presents ambiguous behaviors of the Marathi words, section 7 presents a discussion on special cases, and section 8 concludes the paper with future work. In NLP pipeline POS tagging is an important activity which forms the base of various language processing applications. Annotating a text with POS tags is a standard low-level text pre-processing step before moving to higher levels in the pipeline like chunking, dependency parsing, etc. (Bhattacharyya, 2015) . Identification of the parts of speech such as nouns, verbs, adjectives, adverbs for each word (token) of the sentence helps in analyzing the role of each word in a sentence (Jurafsky D. et. al., 2016) . It represents a token level annotation wherein it assigns a token with POS category.", "cite_spans": [ { "start": 659, "end": 680, "text": "(Bhattacharyya, 2015)", "ref_id": "BIBREF6" }, { "start": 856, "end": 883, "text": "(Jurafsky D. et. al., 2016)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "Aim of POS tagging is to create a large annotated corpora for natural language processing, speech recognition and other related applications. Annotated corpora serve as an important resource in NLP activities. It proves to be a basic building block for constructing statistical models for the automatic processing of natural languages. The significance of large annotated corpora is widely appreciated by researchers and application developers. Various research institutes in India viz., IIT Bombay 4 , IIIT Hyderabad 5 , JNU New Delhi 6 , and other institutes have developed a large corpus of POS tagged data. In Marathi, there is around 100k annotated data developed as a part of Indian Languages Corpora Initiative (ILCI) 7 project funded by MeitY 8 , New Delhi. This ILCI corpus consists of four domains viz., Tourism, Health, Agriculture, and Entertainment. This tagged data (Tourism -25K, Health -25K, Agriculture -10K, Entertainment -10K, General -30K) is used for various applications like chunking, dependency tree banking, word sense disambiguation, etc. This ILCI annotated data forms a baseline for Marathi POS tagging and is available for download at TDIL portal 9 .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Marathi Annotated Corpora", "sec_num": "3" }, { "text": "The Bureau of Indian Standards (BIS) 10 has come up with a standard set of tags for annotating data for Indian languages. This tag-set is prepared for Hindi under the guidance of BIS. The BIS tag-set aims to ensure standardization in the POS tagging across the Indian languages. The tag sets of all Indian languages have been drafted by Dept. of Information Technology, MeitY and presented as Unified POS standard in Indian languages 11 . Marathi POS tag-set has been prepared at IIT Bombay referring to the standard BIS POS Tag-set, IIIT Hyderabad guideline document (Bharati et al, 2006) and Konkani Tag-set (Vaz et. al., 2012) . This Marathi POS Tag-set can be seen in Appendix A.", "cite_spans": [ { "start": 568, "end": 589, "text": "(Bharati et al, 2006)", "ref_id": "BIBREF0" }, { "start": 610, "end": 629, "text": "(Vaz et. al., 2012)", "ref_id": "BIBREF4" } ], "ref_spans": [], "eq_spans": [], "section": "The Marathi POS Tag-Set", "sec_num": "4" }, { "text": "Challenges and Discussions", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Lexical and Functional POS Tagging:", "sec_num": "5" }, { "text": "Lexical POS tagging (Lexical or L approach) deals with tagging of a word at a token level. Functional POS tagging (Functional or F approach) deals with tagging of a word as a syntactic function of a word in a sentence. In other words, a word can have two roles viz., grammatical role (lexical POS w.r.t. a dictionary entry) and functional role (contextual POS) 12 . For example, in the phrase 'golf stick', the POS tag of the word 'golf' could be determined as follows:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Lexical and Functional POS Tagging:", "sec_num": "5" }, { "text": "\uf0b7 Lexically it is a noun as per lexicon. \uf0b7 Functionally it is an adjective as it is a modifier of succeeding noun. In the initial stage of ILCI data annotation, POS tagging was conducted using the lexical approach. However, over a while, POS tagging was done using the functional approach only. The reason is that, by using the lexical approach we do a general tagging, i.e., tagging at a surface level or token level and by using the functional approach we do a specific tagging, i.e., tagging at a semantic level. This eases the annotation process of chunking and parsing in the NLP pipeline.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Lexical and Functional POS Tagging:", "sec_num": "5" }, { "text": "While performing POS annotation, many issues and challenges were encountered, some of which are discussed below. Table 1 lists the occurrences of discussed words in the ILCI corpus.", "cite_spans": [], "ref_spans": [ { "start": 113, "end": 120, "text": "Table 1", "ref_id": "TABREF1" } ], "eq_spans": [], "section": "Lexical and Functional POS Tagging:", "sec_num": "5" }, { "text": "There are three basic types of adverbs. They are time (N_NST), place (N_NST) and manner (RB). Traditionally, adverbs should be tagged as RB. Subordinators are conjunctions which are tagged as CCS.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Subordinators which act as Adverbs", "sec_num": "5.1" }, { "text": "However, there are some subordinators in Marathi which act as adverbs. For example, \u093e \u092e\u093e\u0923\u0947 (jyApramANe, like-wise), \u093e \u092e\u093e\u0923\u0947 (tyApramANe, like that), \u093e \u092e\u093e\u0923\u0947 (hyA-pramANe, like this), \u091c\u0947 \u093e (jevhA, when) and \u0924\u0947 \u093e (tevhA, then). \u093e \u092e\u093e\u0923\u0947 (jyApramANe) and \u093e \u092e\u093e\u0923\u0947 (tyApramANe) are generated from pronominal stems viz., \u093e (jyA) and \u093e (hyA) hence they are lexically qualified as pronouns, however, they function as adverbs; hence to be functionally tagged as RB at the individual level. However, when these words appear as part of the clause then they should be functionally tagged as CCS.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Subordinators which act as Adverbs", "sec_num": "5.1" }, { "text": "This distinction was also observed by noted Marathi grammarian, Damle 13 (Damle, 1965) [p. 206-07].", "cite_spans": [ { "start": 73, "end": 86, "text": "(Damle, 1965)", "ref_id": "BIBREF2" } ], "ref_spans": [], "eq_spans": [], "section": "Subordinators which act as Adverbs", "sec_num": "5.1" }, { "text": "There are suffixes like \u092e\u0941 \u0933\u0947 (muLe, because of; due to), \u0938\u093e\u0920\u0940 (sAThI, for), \u092c\u0930\u094b\u092c\u0930, (barobara, along with), etc. When these suffixes are attached to pronouns they function as adverbs or conjunctions at a syntactic level. For example, words \u093e\u092e\u0941 \u0933\u0947 (tyAmuLe, because of that), \u092f\u093e\u092e\u0941 \u0933\u0947 (yAmuLe, because of this), \u092f\u093e\u0938\u093e\u0920\u0940 (yAsAThI, for this), \u093e\u092e\u0941 \u0933\u0947 (hyAmuLe, because of this), \u093e \u093e\u092e\u0941 \u0933\u0947 (hyAchyA-muLe, because of it/him), \u092f\u093e\u0902 \u093e\u092e\u0941 \u0933\u0947 (yAMchyAmuLe, because of them), \u093e\u091a\u092c\u0930\u094b\u092c\u0930 (=\u0924\u0938\u0947 \u091a) (tyAchabarobara (=tasecha), further) are formed by attaching the above suffixes to pronouns. These words which are formed are lexically tagged as PRP. However, functionally these words act as conjunctions at the sentence level; therefore, they should be tagged as CCD. Also, consider the words \u093e\u0935\u0947 \u0933\u0940 (tyAveLI, at that time), \u093e\u0935\u0947 \u0933\u0940 (hyAveLI, at this time), \u093e\u0928\u0902 \u0924\u0930 (hyAnaMtara, after this), \u093e\u0928\u0902 \u0924\u0930 (tyAnaMtara, after that). Here, wherever the first string/morpheme appears as \u093e (tyA) and \u093e (hyA), the tag should be given as PRP, lexically. But functionally, all these words shall be tagged as N_NST (time adverb).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Words with Suffixes", "sec_num": "5.2" }, { "text": "Adjectives are tagged as JJ. Consider the example below:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Words which are Adjectives", "sec_num": "5.3" }, { "text": "\u093e \u093e\u092e \u0947 \u0939\u0940 \u0915\u0932\u093e \u092a\u0930\u0902 \u092a\u0930\u093e\u0917\u0924 \u091a\u093e\u0932\u0924 \u0906\u0932\u0940 \u0906\u0939\u0947 (tyAchyAmadhye hI kalA paraMparAgata chAlataAlIAhe, this art has come to him by tradition). Lexically, the word \u092a\u0930\u0902 \u092a\u0930\u093e\u0917\u0924 (pa-raMparAgata, traditional) is an adjective, but, in the above sentence, it qualifies the verb \u091a\u093e\u0932\u0924 \u092f\u0947 \u0923\u0947 (chAla-tayeNe, to be practiced). Hence functionally, the word \u092a\u0930\u0902 \u092a\u0930\u093e\u0917\u0924 (paraMparAgata) should be tagged as an RB. Similarly, a word \u0935\u093e\u0908\u091f (vAITa, bad) has a lexical POS as an adjective (Date-Karve, 1932) . But in the sentence \u092e\u0932\u093e \u0935\u093e\u0908\u091f \u0935\u093e\u091f\u0924\u0947 (malA vAITa vATate, I am feeling bad), it functions as an adverb, as it is qualifying the verb and not preceding the pronoun \u092e\u0932\u093e (malA, I; me). Therefore, functionally word \u0935\u093e\u0908\u091f (vAITa) acts as adverb hence should be tagged as RB.", "cite_spans": [ { "start": 452, "end": 470, "text": "(Date-Karve, 1932)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Words which are Adjectives", "sec_num": "5.3" }, { "text": "The adnominal suffix \u091c\u094b\u0917\u0902 (jogaM) and all its forms (\u091c\u094b\u0917\u093e, \u091c\u094b\u0917\u0940, \u091c\u094b\u0917\u0947 , \u091c\u094b \u093e; jogA, jogI, joge, jogyA) are always attached to verbs. For example, word \u0915\u0930 \u093e\u091c\u094b \u093e (karaNyA-jogyA, doable) is lexically tagged as a verb. However, word \u0915\u0930 \u093e (karaNyA) is a Kridanta form of a verb \u0915\u0930\u0923\u0947 (karaNe, to do) and suffix \u091c\u094b\u0917\u0902 (jogaM) is an adnominal suffix attached to Kridanta form; hence, a verb with all the forms of \u091c\u094b\u0917\u0902 (jogaM) should functionally be treated as adjectives. Therefore verbs with adnominal suffix should be tagged as JJ.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Adnominal Suffixes Attached to Verbs", "sec_num": "5.4" }, { "text": "As per Damle (1956) , words \u091c\u0938\u0947 (jase, like this) and \u0924\u0938\u0947 (tase, like that) are tagged as adverbs. However, if they appear with nouns in a sentence, they are influenced by the inflection and gender property of that nominal stem. For example, words \u091c\u0938\u0947 (jase, like this) and \u0924\u0938\u0947 (tase, like that) have inflected forms like \u091c\u0938\u093e (jasA, like him), \u091c\u0936\u0940 (jashI, like her), \u0924\u0938\u093e (tasA, like this), \u0924\u0936\u0940 (tashI, like this), \u0924\u0938\u0947 (tase, like this), etc. All these words function as a relative pronoun in a sentence. Hence, the words and their variations should be functionally tagged as PRL.", "cite_spans": [ { "start": 7, "end": 19, "text": "Damle (1956)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Words \u091c\u0938\u0947 (jase) \u0924\u0938\u0947 (tase)", "sec_num": "5.5" }, { "text": "A word \u0924\u0938\u0947 \u0924\u0930 (tasetara, as it is seen) is the same as \u0924\u0938\u0947 \u092a\u093e\u093f\u0939\u0932\u0947 \u0924\u0930 (tase pAhile tara, as it is seen). Lexically, it can be tagged as a particle (RPD) but since it has a function of conjunction; it should be tagged as CCD. For example, in a sentence \u0924\u0938\u0947 \u0924\u0930 \u0924\u0923\u093e\u0935\u093e\u092e\u0941 \u0933\u0947 \u0939\u0940 \u0915\u093e\u0933\u0940 \u0935\u0932\u092f \u092f\u0947 \u0924\u093e\u0924 (tasetara ta-NAvAmuLehI kALI valaya yetAta, as it is seen that black circles appear because of stress as well), word \u0924\u0938\u0947 \u0924\u0930 (tasetara) functions as conjunction and hence should be tagged as CCD instead of tagging it as RPD.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word \u0924\u0938\u0947 \u0924\u0930 (tasetara)", "sec_num": "5.6" }, { "text": "The standard dictionaries give POS of the word \u0905 \u0925\u093e (anyathA, otherwise; else; or) as an adverb/indeclinable. For example, consider a sentence \u0905 \u0925\u093e \u0924\u094b \u092f\u0947 \u0923\u093e\u0930 \u0928\u093e\u0939\u0940 (any-athA to yeNAra nAhI, Otherwise, he will not come). Here, while annotating \u0905 \u0925\u093e (anyathA) there is a possibility that annotator can directly tag this word as an adverb at a lexical level. However, it behaves like conjunction at the sentence level and hence it should be tagged as CCD.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word \u0905 \u0925\u093e (anyathA)", "sec_num": "5.7" }, { "text": "As per BIS Tag-set, words \u0915\u0938\u093e, \u0915\u0936\u0940, \u0915\u0938\u0947 (kasA, kashI, kase; how) shall be tagged as PRQ. However, the PRQ tag is only for pronoun category and the word \u0915\u0938\u093e (kasA) is not a pronoun; it can behave as an adverb or as a modifier. Consider the examples below:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Different Forms of \u0915\u0938\u093e (kasA)", "sec_num": "5.8" }, { "text": "1. \u0924\u094b \u092e\u093e\u0923\u0942 \u0938 \u0915\u0938\u093e \u0906\u0939\u0947 \u0939\u0947 \u093e \u093e\u0936\u0940 \u092c\u094b\u0932 \u093e\u0935\u0930\u091a \u0915\u0933\u0947 \u0932 (to mANUsa kasA Ahe he tyAchyAshI bolalyAvaracha kaLela, we will come to know about him only after talking to him) [adnominal] 2. \u0938\u0930\u0915\u093e\u0930\u0940 \u0920\u0930\u093e\u0935\u093e\u0928\u0947 \u0915\u093e\u092f \u093e\u091a\u0947 \u0915\u0932\u092e \u0915\u0938\u0947 \u0930 \u0939\u094b\u0923\u093e\u0930 (sarakArI TharAvAne kAyadyAche kalama kase radda hoNAra, How can this clause of law be prohibited by Government Resolution?) [adverbial] In the 1st case, word \u0915\u0938\u093e (kasA, how) functionally acts as a pronoun, hence to be tagged as PRQ. While, in the 2nd case, it acts as an adverb, hence to be functionally tagged as RB.", "cite_spans": [ { "start": 339, "end": 350, "text": "[adverbial]", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Different Forms of \u0915\u0938\u093e (kasA)", "sec_num": "5.8" }, { "text": "A word \u092e\u093e (mAtra) is very ambiguous in its various usages; it is difficult to functionally identify the POS of this word at a sentence level. Various meanings of word \u092e\u093e (mAtra) are given in Data-Karve dictionary 14 . Some of the different senses of \u092e\u093e (mAtra) are discussed here:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word \u092e\u093e (mAtra)", "sec_num": "5.9" }, { "text": "\uf0b7 When the word \u092e\u093e (mAtra) conveys the meaning of \u0939\u0940, \u0926\u0947 \u0916\u0940\u0932, \u0938\u0941 \u093e (hI, dekhIla, suddhA; also) then it should be tagged as RB functionally. \uf0b7 When a word is related to the preceding word \u0924\u0947 \u0925\u0947 (tethe, there) and its function is an emphatic marker \u091a (cha) then it should be tagged as RPD functionally. \uf0b7 When word \u092e\u093e (mAtra) appears in the form of conjunction then it should be marked as CC functionally. \uf0b7 If the word is modifying the succeeding noun, then it should be tagged as JJ functionally. \uf0b7 If the word is modifying the preceding word, then the tag will be RPD as a particle functionally.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word \u092e\u093e (mAtra)", "sec_num": "5.9" }, { "text": "Therefore, it is noticed that the word \u092e\u093e (mAtra) does not have one single POS tag functionally and it depends upon the appearance in a sentence. Hence, should be tagged as per the usage. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Word \u092e\u093e (mAtra)", "sec_num": "5.9" }, { "text": "Examples: \u0935\u092f \u0930 (vayaskara, the aged) \uf0b7 \u0915 \u0941 \u091f\u0941 \u0902 \u092c\u093e \u093e \u0935\u092f \u0930 \u0938\u0926 \u093e\u0902 \u0928\u0940 \u092e\u0924\u0926\u093e\u0928 \u0915 \u0947 \u0932\u0947 (kuTuM-bAchyA vayaskara sadasyAMnI matadAnakele, all the aged members of the family voted). \uf0b7 \u0938\u0935 \u0935\u092f \u0930\u093e\u0902 \u0928\u0940 \u092e\u0924\u0926\u093e\u0928 \u0915 \u0947 \u0932\u0947 (sarva vayaskarAMnI matadAna kele, all the aged people voted).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Adjective or Noun?", "sec_num": "6.1" }, { "text": "In the above examples, the word \u0935\u092f \u0930 (vayaskarAMnI) lexically acts as an adjective as well as a noun. However, at the syntactic level, in the first example, it is functioning as adjective hence to be tagged as JJ, while in the second example it is functioning as a noun hence to be tagged as N_NN. This is one of the challenges while annotating adjectives appearing in nominal form. Annotators usually fail to disambiguate these types of words at the lexical level; therefore such words should be disambiguated at syntactic level. Hence, annotators need to take special care while annotating such cases.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Adjective or Noun?", "sec_num": "6.1" }, { "text": "While annotating demonstrators such as \u0939\u093e, \u0939\u0940, \u0939\u0947 , \u0924\u094b, \u0924\u0940, \u0924\u0947 ((hA, hI, he), this), (to, tI, te), that) annotators often get confused whether to tag them as DMD or DMR. Simple guideline can be followed is, if the demonstrator is directly following noun, then tag it as DMD, otherwise tag it as DMR i.e., if the demonstrator is referring to previous noun/person.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Demonstrators", "sec_num": "6.2" }, { "text": "Example: word \u0915\u093e\u0930\u0923 (kAraNa, reason; because). At semantic level, the word \u0915\u093e\u0930\u0923 (kAraNa) has two meanings, one is 'a reason' which acts as a noun and another is 'because' which acts as a conjunction. Annotators have to pay special attention while tagging such cases.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "6.4 Ambiguous Words: \u0924\u0947 (te) and \u0924\u0947 \u0939\u0940 (tehI)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "The word \u0924\u0947 (te) has different grammatical categories like pronoun (they), demonstrator (that) and conjunction (to). Examples:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 \u0969\u0966 \u0924\u0947 \u096a\u0966 (30 te 40, 30 to 40) \uf0b7 The word \u0924\u0947 (te) lexically and functionally acts as conjunction, hence to be tagged as CCD. \uf0b7 \u0924\u0947 \u0923\u093e\u0932\u0947 (te mhaNAle, they said) \uf0b7 Here word \u0924\u0947 (te) acts as personal pronoun, hence to be tagged as PR_PRP \uf0b7 \u0924\u0947 \u0915 \u0941 \u0920\u0947 \u0906\u0939\u0947 \u0924? (te kuThe Aheta?, where are they?) \uf0b7 Here word \u0924\u0947 (te) acts as relative demonstrator, hence to be tagged as DM_DMR \uf0b7 \u0930\u093e\u0915 \u0947 \u0936\u0928\u0947 \u092a\u094b\u0932\u0940\u0938\u093e\u0902 \u0928\u093e \u092b\u094b\u0928 \u0915 \u0947 \u0932\u093e \u0906\u093f\u0923 \u0924\u0947 \u0926\u094b \u0940 \u091a\u094b\u0930 \u092a\u0915\u0921\u0932\u0947 \u0917\u0947 \u0932\u0947 (rAkeshane polIsAMnA phona kelA ANi te donhI chora pakaDale gele, Rakesh called police and those two thieves got caught). \uf0b7 Here, word \u0924\u0947 (te) is modifying its succeeding noun \u091a\u094b\u0930 (chora, thief) so it is Deictic demonstrator, hence to be tagged as DM_DMD.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 \u093e\u0902 \u0928\u093e \u0939\u0947 \u0915\u0927\u0940\u091a \u092a\u0938\u0902 \u0924 \u0928 \u0924\u0947 , \u093e\u0902 \u093e \u092e\u0941 \u0932\u093e\u0928\u0947 \u0938\u0902 \u0917\u0940\u0924 \u093f\u0936\u0915\u093e\u0935\u0947 \u0906\u093f\u0923 \u0924\u0947 \u0939\u0940 \u0928\u0943 (tyAMnA he kadhIchapasaMtanavhate, tyAMchyAmulAnesaMgItashikAveANite-hInRRitya, He never wanted his son to learn music and that too the dance form)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Here, the word \u0924\u0947 \u0939\u0940 (tehI) is an ambiguous word. It is modifying succeeding noun or previous context. Here, \u0939\u0940 (hI) is a bound morpheme and conveys the meaning 'also'. Therefore word \u0924\u0947 \u0939\u0940 (tehI) should be tagged as DM_DMR.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "6.5 Ambiguous word: \u0909\u0932\u091f\u093e (ulaTA)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Examples:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 \u0909\u0932\u091f\u0947 \u091f\u093e\u0902 \u0917\u0942 \u0928 \u0938\u0941 \u0915\u0935\u0932\u0947 \u091c\u093e\u0924\u0947 (ulaTe TAMgUna sukavale jAte). Here, \u0909\u0932\u091f\u0947 (ulaTe, upside down is behaving as manner, not a noun, hence to be tagged as RB. \uf0b7 \u0909\u0932\u091f\u0947 \u092d\u093e\u0902 \u0921\u0947 \u0938\u0941 \u0932\u091f\u0947 \u0915\u0930 (ulaTe bhAMDe sulaTe kara).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Here \u0909\u0932\u091f\u0947 (ulaTe) it is modifying succeeding noun, hence it is an adjective, hence to be tagged as JJ.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "In the above examples, annotator should identify word behavior in the sentence and tag accordingly.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "6.6 Ambiguous words: \u093f\u0915\u0924\u0940\u0939\u0940 (kitIhI), \u0928\u093e \u0915\u093e (nA kA) and \u0905\u0938\u0942 \u0926\u0947 \u0928\u093e \u0915\u093e (asU de nA kA)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Examples:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 \u0938\u0902 \u0917\u0923\u0915 \u0939\u093e \u093f\u0915\u0924\u0940\u0939\u0940 \u0917\u0924 \u093f\u0915 \u0902 \u0935\u093e \u091a\u0924\u0941 \u0930 \u0905\u0938\u0942 \u0926\u0947 \u0928\u093e \u0915\u093e, \u0924\u094b \u0915 \u0947 \u0935\u0933 \u0924\u0947 \u091a \u0915\u093e\u092e \u0915 \u0936\u0915\u0924\u094b \u093e\u091a\u0940 \u093f\u0935\u0927\u0940 (\u092a \u0924) \u0906\u092a \u093e\u0932\u093e \u0924: \u092e\u093e\u093f\u0939\u0924 \u0906\u0939\u0947 . (saMgaNaka hA kitIhI pragata kiMvA chatura asU de nA kA, to kevaLa techa kAma karU shakato jyAchI vidhI (paddhata) ApalyAlA svata: mAhita Ahe\u00b8 The computer how much ever may be advanced and clever, it only does that work whose method we only know). Here, \u093f\u0915\u0924\u0940\u0939\u0940 (kitIhI, how much) is a quantifier, hence to be tagged as QTF. \uf0b7 In the phrase \u0905\u0938\u0942 \u0926\u0947 \u0928\u093e \u0915\u093e (asU de nA kA), the token \u0928\u093e (nA) is a part of verb \u0905\u0938\u0942 \u0926\u0947 (asU de, let it be) and should be tagged as VM, hence the phrase should be tagged as VM, while the token \u0915\u093e (kA) is acting as a particle in this phrase and not as a question marker, therefore \u0915\u093e (kA) should be tagged as RPD. \uf0b7 \u093f\u0915\u0924\u0940 \u092e\u093e\u0923\u0938\u0947 \u091c\u0947 \u0935\u093e\u092f\u0932\u093e \u0939\u094b\u0924\u0940? (kitI mANase jevAyalA hotI, how many people were there for a meal?).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Here, \u093f\u0915\u0924\u0940 (kitI, how many) is a question so it should be tagged as DMQ.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "6.7 Ambiguous word: \u0924\u0930 (tara)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Examples:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 Conjunction: \u091c\u0930 \u092e\u0940 \u0935\u0947 \u0933\u0940\u091a \u0917\u0947 \u0932\u094b \u0928\u0938\u0924\u094b \u0924\u0930 \u0939\u093e \u0935\u093e\u091a\u0932\u093e \u0928\u0938\u0924\u093e (jara mI veLIcha gelo nasato tara hA vA-chalA nasatA, if I had not gone on time he would have not survived). \uf0b7 Particle: '\u0939\u094b! \u0906\u0924\u093e \u092e\u0940 \u091c\u093e\u0924\u094b \u0924\u0930!' = '\u092e\u0940 \u0905\u093f\u091c\u092c\u093e\u0924 \u091c\u093e\u0923\u093e\u0930 \u0928\u093e\u0939\u0940'('ho! AtA mI jAto tara!' = 'mI ajibAta jANA-ra nAhI', 'yes! now I am leaving then' = 'I am not at all leaving').", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "In the above sentences, the word \u0924\u0930 (tara) is used as a supplementary or stressable word so somewhat special as to give meaning in the sentence. (Date-Karve, 1932) . Hence it should be treated as CCD.", "cite_spans": [ { "start": 145, "end": 163, "text": "(Date-Karve, 1932)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "\uf0b7 \u0924\u0941 \u0940 \u0924\u0930 \u0932\u093e\u0916 \u092a\u092f\u0947 \u092e\u093e\u0917\u0924\u093e\u0902 \u0935 \u092e\u0940 \u0924\u0930 \u0915 \u0947 \u0935\u0933 \u0917\u0930\u0940\u092c \u092a\u0921\u0932\u094b (tumhI tara lAkha rupaye mAgatAM va mI tara kevaLa garIba paDalo, you are asking for lakh rupees and I am a poor person). In this sentence, the word \u0924\u0930 (tara) indicates opposition with respect to meaning between two connected sentences. (Date-Karve, 1932) . Hence, it should be treated as a RPD.", "cite_spans": [ { "start": 287, "end": 305, "text": "(Date-Karve, 1932)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "7 Discussions on Some Special Cases \uf0b7 Words \u0906 \u092f\u0941 (Amlayukta), \u092e\u0932\u0908\u0930\u093f\u0939\u0924 (malaIrahita), \u092e\u0947 \u0926\u0930\u093f\u0939\u0924 (medarahita), \u0926\u0941 \u093e\u0933 (duSh-kALagrasta) are combinations of noun plus adjective suffix such as \u092f\u0941 (yukta), (grasta) and \u0930\u093f\u0939\u0924 (rahita). In such cases, even though noun is a head string and adjective part is a suffix, the whole word shall be tagged as JJ. \uf0b7 Before tagging \u0905\u092d\u0902 \u0917 (abhaMga, verses), \u0913 \u093e (ovyA, stanzas), \u0915\u093e (kAvya, poetry), etc., annotator shall first read between the lines; understand the meaning which it conveys and then decide upon the grammatical categories of each token. Here, in the 1st example, the word \u0915\u094b\u0923\u0924\u093e (koNa-tA, which one) undoubtedly is DMQ. In rest of the examples \u0915\u094b\u0923\u0924\u0940\u0939\u0940 (koNatAhI, whichever, whomever), \u0915\u094b\u0923\u0924\u093e\u0939\u0940 (koNatAhI, whichever, whomever), \u0915\u094b\u0923 \u093e\u0939\u0940 (koNatyAhI, whichever, whomever) are DM adjective (DMD).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Ambiguous POS: Noun and Conjunction", "sec_num": "6.3" }, { "text": "Marathi POS tagging is an important activity for NLP tasks. While tagging, several challenges and issues were encountered. In this paper, Marathi BIS tag-set has been discussed. Lexical and functional tagging approaches were discussed with examples. Further, various challenges, experiences, and special cases have been presented. The issues discussed here will be helpful for annotators, researchers, language learners, etc. of Marathi and other languages.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Work", "sec_num": "8" }, { "text": "In future, more issues such as tagging for words having multiple senses; words having multiple functional tags will be discussed. Also, tagset comparison of close languages will be done. Further, the evaluation of lexical and functional tagging using statistical analysis will be done. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion and Future Work", "sec_num": "8" }, { "text": "http://www.cfilt.iitb.ac.in/ 5 https://ltrc.iiit.ac.in/ 6 https://www.jnu.ac.in/ 7 http://sanskrit.jnu.ac.in/ilci/index.jsp 8 https://meity.gov.in/", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "https://www.tdil-dc.in/ 10 http://www.bis.gov.in/ 11 http://tdil-dc.in/tdildcMain/articles/134692Draft POS Tag standard.pdf 12 https://www.cicling.org/2018/cicling18-hanoi-special-event-23mar18.pdf", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "http://www.cfilt.iitb.ac.in/damale/index.html", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "http://www.transliteral.org/dictionary/mr.kosh.maharasht ra/source", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null } ], "back_matter": [], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "AnnCorra : Annotating Corpora Guidelines For POS And Chunk Annotation For Indian Languages", "authors": [ { "first": "Akshar", "middle": [], "last": "Bharati", "suffix": "" }, { "first": "Dipti", "middle": [ "Misra" ], "last": "Sharma", "suffix": "" }, { "first": "Lakshmi", "middle": [], "last": "Bai", "suffix": "" }, { "first": "Rajeev", "middle": [], "last": "Sangal", "suffix": "" } ], "year": 2006, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Akshar Bharati, Dipti Misra Sharma, Lakshmi Bai, Rajeev Sangal. (2006). AnnCorra : Annotating Corpora Guide- lines For POS And Chunk Annotation For Indian Lan- guages. Language Technologies Research Centre, IIIT, Hyderabad.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Sentiment Analysis in Marathi using Marathi WordNet", "authors": [ { "first": "V", "middle": [], "last": "Chitra", "suffix": "" }, { "first": "Ashwini", "middle": [ "V" ], "last": "Chaudhari", "suffix": "" }, { "first": "Rashmi", "middle": [ "R" ], "last": "Khaire", "suffix": "" }, { "first": "Komal", "middle": [ "S" ], "last": "Murtadak", "suffix": "" }, { "first": "", "middle": [], "last": "Sirsulla", "suffix": "" } ], "year": 2017, "venue": "Imperial Journal of Interdisciplinary Research (IJIR)", "volume": "3", "issue": "", "pages": "2454--1362", "other_ids": {}, "num": null, "urls": [], "raw_text": "Chitra V. Chaudhari, Ashwini V. Khaire, Rashmi R. Mur- tadak, Komal S. Sirsulla. (2017). Sentiment Analysis in Marathi using Marathi WordNet. Imperial Journal of Interdisciplinary Research (IJIR) Vol-3, Issue-4, 2017 ISSN: 2454-1362.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Shastriya Marathi Vyakran. A scientific grammar of Marathi", "authors": [ { "first": "Moro", "middle": [], "last": "Damle", "suffix": "" }, { "first": "", "middle": [], "last": "Keshav", "suffix": "" } ], "year": 1965, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Damle, Moro Keshav. (1965). Shastriya Marathi Vyakran. A scientific grammar of Marathi, 3 rd edition. Pune, In- dia: RD Yande.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Speech and Language Processing", "authors": [ { "first": "Daniel", "middle": [], "last": "Jurafsky", "suffix": "" }, { "first": "&", "middle": [], "last": "James", "suffix": "" }, { "first": "H", "middle": [], "last": "Martin", "suffix": "" } ], "year": 2016, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Daniel Jurafsky & James H. Martin. (2016). Speech and Language Processing.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "BIS Annotation Standards With Reference to Konkani Language. 24 th International Conference on Computational Linguistics (COL-ING 2012)", "authors": [ { "first": "Edna", "middle": [], "last": "Vaz", "suffix": "" }, { "first": "V", "middle": [], "last": "Shantaram", "suffix": "" }, { "first": "Dr", "middle": [ "Jyoti" ], "last": "Walawalikar", "suffix": "" }, { "first": "Dr", "middle": [ "Madhavi" ], "last": "Pawar", "suffix": "" }, { "first": "", "middle": [], "last": "Sardesai", "suffix": "" } ], "year": 2012, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Edna Vaz, Shantaram V. Walawalikar, Dr. Jyoti Pawar, Dr. Madhavi Sardesai. (2012). BIS Annotation Standards With Reference to Konkani Language. 24 th Interna- tional Conference on Computational Linguistics (COL- ING 2012), Mumbai.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Creating Marathi WordNet. The WordNet in Indian Languages", "authors": [ { "first": "Lata", "middle": [], "last": "Popale", "suffix": "" }, { "first": "Pushpak", "middle": [], "last": "Bhattacharyya", "suffix": "" } ], "year": 2017, "venue": "", "volume": "", "issue": "", "pages": "147--166", "other_ids": {}, "num": null, "urls": [], "raw_text": "Lata Popale and Pushpak Bhattacharyya. (2017). Creating Marathi WordNet. The WordNet in Indian Languages. Springer, Singapore, 2017. 147-166.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Machine Translation", "authors": [ { "first": "Pushpak", "middle": [], "last": "Bhattacharyya", "suffix": "" } ], "year": 2015, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Pushpak Bhattacharyya, (2015). Machine Translation, Book published by CRC Press, Taylor and Francis Group, USA.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Published by H. A. Bhave, Varada Books, Senapati Bapat Marg", "authors": [ { "first": "Maharashtra", "middle": [], "last": "Shabdakosh", "suffix": "" } ], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Maharashtra Shabdakosh. Published by H. A. Bhave, Varada Books, Senapati Bapat Marg, Pune.", "links": null } }, "ref_entries": { "FIGREF0": { "num": null, "type_str": "figure", "text": "1 http://language.worldofcomputing.net/pos-tagging/parts-ofspeech-tagging.html 2 http://www.indianmirror.com/languages/marathi-language.html 3 http://www.censusindia.gov.in/ 2 Parts-Of-Speech Annotation", "uris": null }, "TABREF1": { "content": "
: Occurrences of discussed words and lexical v/s
functional tags assigned to these words
6 POS Ambiguity: Challenges and Discus-
sions
", "num": null, "text": "", "html": null, "type_str": "table" }, "TABREF2": { "content": "
Examples:
o \u0915\u094b\u0923\u0924\u093e \u092e\u0941 \u0932\u0917\u093e \u0936\u093e\u0930 \u0906\u0939\u0947 (koNatA mulagA hu-
shAra Ahe)?
o \u0935\u093e\u0939\u0924\u0941 \u0915\u0940 \u093e \u0926\u0930 \u093e\u0928 \u0915\u094b\u0923\u0924\u0940\u0939\u0940 \u0939\u093e\u0928\u0940 \u091d\u093e\u0932\u0947 \u0932\u0940 \u0928\u093e\u0939\u0940
(vAhatukIchyA daramyAna koNatIhI hAnI
jhAlelI nAhI).
o\u093e\u0902 \u093e \u092c\u094b\u0932 \u093e\u091a\u093e \u092e\u093e \u093e\u0935\u0930 \u0915\u094b\u0923\u0924\u093e\u0939\u0940 \u092a \u0930\u0923\u093e\u092e \u091d\u093e\u0932\u093e
\u0928\u093e\u0939\u0940 (hyAMchyA bolaNyAchA mAjhyAvara
koNatAhI pariNAma jhAlA nAhI).
o \u0936\u0947 \u0924\u0915 \u092f\u093e\u0938 \u0915\u094b\u0923 \u093e\u0939\u0940 \u0935\u0937 \u092a\u093e \u093e\u091a\u0940 \u0915\u092e\u0924\u0930\u0924\u093e
\u092d\u093e\u0938\u0923\u093e\u0930 \u0928\u093e\u0939\u0940 (shetakaryAsa koNatyAhI var-
ShI pANyAchI kamataratA bhAsaNAra nA-
hI).
", "num": null, "text": "For example, in sentence \u0915\u0933\u093e\u0935\u0947 \u0924\u092f\u093e\u0938\u0940 \u0915\u0933\u0947 \u0905\u0902 \u0924\u0930\u0940\u091a\u0947 \u0915\u093e\u0930\u0923 \u0924\u0947 \u0938\u093e\u091a\u0947 \u0938\u093e\u091a \u0905\u0902 \u0917\u0940 (kaLAve tayAsI kale aMtarIche kAraNa te sAche sAcha aMgI) the POS tagging should be done as \u0938\u093e\u091a\u0947 \\V_VM \u0938\u093e\u091a\\N_NN \u0905\u0902 \u0917\u0940\\N_NN, etc. \uf0b7 Doubtful cases of word \u0915\u094b\u0923\u0924\u093e (koNatA)", "html": null, "type_str": "table" }, "TABREF3": { "content": "
Appendix A
Marathi Parts of Speech Tag-Set with Examples
SI. NoCategoryLabelAnnotation ** ConventionExamples
Top level & Subtype
1Noun (\u0928\u093e\u092e)NN
1.1Common (\u091c\u093e\u0924\u0940\u0935\u093e\u091a\u0915 \u0928\u093e\u092e)NNN_NN\u0917\u093e\u092f\\N_NN \u0917\u094b \u093e\u0924\\N_NN \u0930\u093e\u0939\u0924\u0947 .
1.2Proper (\u0940\u0935\u093e\u091a\u0915 \u0928\u093e\u092e)NNPN_NNP\u0930\u093e\u092e\u093e\u0928\u0947 \\N_NNP \u0930\u093e\u0935\u0923\u093e\u0932\u093e\\N_NNP \u092e\u093e\u0930\u0932\u0947 .
1.3 1. 2 Nloc ( \u0925\u0932-\u0915\u093e\u0932) NST N_NST Pronoun (\u0938\u0935 \u0928\u093e\u092e) PR PR
2.1Personal (\u092a\u0941 \u0937 \u0935\u093e\u091a\u0915)PRPPR_PRP\u092e\u0940\\PR_PRP \u092f\u0947 \u0924\u094b.
2.2Reflexive (\u0906 \u0935\u093e\u091a\u0915)PRFPR_PRF\u092e\u0940 \u0924\u0903\\PR_PRF \u0906\u0932\u094b.
2.3Relative (\u0938\u0902 \u092c\u0902 \u0927\u0940)PRLPR_PRL\u093e\u0928\u0947 \\PR_PRL \u0939\u0947 \u0938\u093e\u0902 \u093f\u0917\u0924\u0932\u0947 \u093e\u0928\u0947 \u0939\u0947 \u0915\u093e\u092e \u0915 \u0947 \u0932\u0947 \u092a\u093e\u093f\u0939\u091c\u0947 .
2.4Reciprocal (\u092a\u093e\u0930 \u0930\u0915)PRCPR_PRC\u092a\u0930 \u0930
2.5Wh-word ( \u093e\u0925 \u0915)PRQPR_PRQ\u0915\u094b\u0923\\PR_PRQ \u092f\u0947 \u0924 \u0906\u0939\u0947 ?
2.6Indefinite (\u0905\u093f\u0928\u093f \u0924)PRIPR_PRI
3Demonstrative (\u0926\u0936 \u0915)DMDM\u0939\u0947 \u092a\u0941 \u0915 \u092e\u093e\u091d\u0947 \u0906\u0939\u0947 .
3.1DeicticDMDDM_DMD
3.2RelativeDMRDM_DMR\u0939\u0947 \\DM_DMR \u0932\u093e\u0932 \u0930\u0902 \u0917\u093e\u091a\u0947 \u0905\u0938\u0924\u0947 .
3.3Wh-wordDMQDM_DMQ\u0915\u094b\u0923\u0924\u093e\\DM_DMQ \u092e\u0941 \u0932\u0917\u093e \u0936\u093e\u0930 \u0906\u0939\u0947 ?
4Verb (\u093f \u092f\u093e\u092a\u0926)VV
4.1Main (\u092e\u0941 \u093f \u092f\u093e\u092a\u0926)VMV_VM\u0924\u094b \u0918\u0930\u0940 \u0917\u0947 \u0932\u093e\\V_VM.
4.2Auxiliary (\u0938\u0939\u093e \u0915 \u093f \u092f\u093e\u092a\u0926VAU XV_VAUX\u0930\u093e\u092e \u0918\u0930\u0940 \u091c\u093e\u0924 \u0906\u0939\u0947 \\V_VAUX.
5Adjective (\u093f\u0935\u0936\u0947 \u0937\u0923)JJ\u0938\u0941 \u0902 \u0926\u0930\\JJ \u092e\u0941 \u0932\u0917\u0940
6Adverb (\u093f \u092f\u093e\u093f\u0935\u0936\u0947 \u0937\u0923)RB\u0939\u0933\u0942 \u0939\u0933\u0942 \\RB \u091a\u093e\u0932.
7Conjunction (\u0909\u092d\u092f\u093e \u092f\u0940 \u0905 \u092f)CCCC
7.1CoordinatorCCDCC_CCD\u0924\u094b \u0906\u093f\u0923\\ CC_CCD \u092e\u0940.
7.2SubordinatorCCSCC_CCS\u091c\u0930\\CC_CCS \u093e\u0928\u0947 \u0938\u093e\u0902 \u093f\u0917\u0924\u0932\u0947 \u0905\u0938\u0924\u0947 \u0924\u0930\\CC_CCS \u0939\u0947 \u0915\u093e\u092e \u092e\u0940 \u0915 \u0947 \u0932\u0947 \u0905\u0938\u0924\u0947 .
7.2.1QuotativeUTCC_CCS_U T
8ParticlesRPRP
8.1DefaultRPDRP_RPD\u092e\u0940 \u0924\u0930\\RP_RPD \u0916\u0942 \u092a \u0926\u092e\u0932\u0947 .
8.2Interjection (\u0909 \u093e\u0930 \u0935\u093e\u091a\u0915)INJRP_INJ\u0905\u0930\u0947 \u0930\u0947 \\RP_INJ ! \u0938\u093f\u091a\u0928\u091a\u0940 \u093f\u0935\u0915 \u0947 \u091f \u0922\u093e\u092a\u0932\u0940.
8.3Intensifier (\u0924\u0940 \u0935\u093e\u091a\u0915)INTFRP_INTF\u0930\u093e\u092e \u0916\u0942 \u092a\\RP_INTF \u091a\u093e\u0902 \u0917\u0932\u093e \u092e\u0941 \u0932\u0917\u093e \u0906\u0939\u0947 .
8.4Negation (\u0928\u0915\u093e\u0930\u093e \u0915)NEGRP_NEG\u0928\u0915\u094b, \u0928
9QuantifiersQTQT
9.1GeneralQTFQT_QTF\u0925\u094b\u0921\u0940\\QT_QTF \u0938\u093e\u0916\u0930 \u093e.
9.2CardinalsQTCQT_QTC\u092e\u0932\u093e \u090f\u0915\\QT_QTC \u0917\u094b\u0933\u0940 \u0926\u0947 .
9.3OrdinalsQTOQT_QTO\u092e\u093e\u091d\u093e \u092a\u093f\u0939\u0932\u093e\\QT_QTO \u092e\u093e\u0902 \u0915 \u0906\u0932\u093e.
10Residuals (\u0909\u0935 \u0930\u0924)RDRD
10.1Foreign wordRDFRD_RDF
10.2SymbolSYMRD_SYE$, &, *, (, ),
10.3PunctuationPUN CRD_PUNC. (period), ,(comma), ;(semi-colon), !(exclama-tion),? (question), : (colon), etc.
10.4UnknownUNKRD_UNKNot able to identify the Tag.
10.5Echo-wordsECHRD_ECH
", "num": null, "text": "\u0924\u094b \u092f\u0947 \u0925\u0947 \\N_NST \u0915\u093e\u092e \u0915\u0930\u0924 \u0939\u094b\u0924\u093e. 2. \u093e\u0928\u0947 \u0939\u0940 \u0935 \u0942 \u0916\u093e\u0932\u0940\\N_NST \u0920\u0947 \u0935\u0932\u0940 \u0906\u0939\u0947 . \u0915\u094b\u0923\u0940\\PR_PRI \u0915\u094b\u0923\u093e\u0938\\PR_PRI \u0939\u093e\u0938\u0942 \u0928\u092f\u0947 . \u093e \u092a\u0947 \u091f\u0940\u0924 \u0915\u093e\u092f\\PR_PRI \u0906\u0939\u0947 \u0924\u0947 \u0938\u093e\u0902 \u0917\u093e. \u0924\u094b\\DM_DMD \u092e\u0941 \u0932\u0917\u093e \u0936\u093e\u0930 \u0906\u0939\u0947 . \u0939\u093e\\DM_DMD \u092e\u0941 \u0932\u0917\u093e \u0936\u093e\u0930 \u0906\u0939\u0947 . \u0939\u0940\\DM_DMD \u092e\u0941 \u0932\u0917\u0940 \u0938\u0941 \u0902 \u0926\u0930\u0906\u0939\u0947 . \u091c\u0947 \u0925\u0947 \\DM_DMD \u0930\u093e\u092e \u0939\u094b\u0924\u093e \u0924\u0947 \u0925\u0947 \\DM_DMD \u0924\u094b \u0939\u094b\u0924\u093e. \u0905\u0938\u0947 \\CC_CCS_UT \u0923\u0942 \u0928\\C_CCS_UT \u0924\u094b \u092a\u0941 \u0922\u0947 \u0917\u0947 \u0932\u093e.", "html": null, "type_str": "table" } } } }