Machine Translation

13.1 Language divergences and typology

1. What are linguistic universals and statistical universals?
   Universals hold for every language (e.g. words for people, eating and drinking; nouns and verbs; ways to ask questions or give commands); statistical universals hold for most languages.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 277
2. What are translation divergences, and which two kinds are distinguished?
   Differences between languages that make translation hard: idiosyncratic lexical differences handled one by one (the word for “dog”), and systematic differences that can be modeled generally (verb before or after the object).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 278
3. What is linguistic typology, and what resource catalogs typological facts?
   The study of systematic cross-linguistic similarities and differences; WALS, the World Atlas of Language Structures.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 278
4. What are SVO, SOV and VSO languages? Give examples.
   The basic order of subject, verb and object: SVO (English, German, French, Mandarin), SOV (Hindi, Japanese), VSO (Irish, Arabic).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 278
5. What correlation holds between verb–object order and adpositions?
   VO languages generally have prepositions; OV languages generally have postpositions.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 278
6. Give an example of adjective–noun order differing between languages.
   English puts adjectives before nouns (green witch); Spanish and Modern Hebrew put them after (bruja verde).
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 279
7. Give examples of one-to-many lexical divergences in translation.
   English bass → Spanish lubina (fish) or bajo (instrument); wall → German Wand (inside) or Mauer (outside); brother → Mandarin gege (older) or didi (younger). This links MT to word sense disambiguation.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 279
8. How can grammatical constraints on word choice differ between languages?
   One language may require information another leaves unmarked: English marks plural on nouns but Mandarin doesn’t; French and Spanish mark gender on adjectives, so translating into them requires choosing a gender.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 279
9. Give an example of a many-to-many lexical mapping between languages.
   English leg, foot, paw vs French: an animal’s leg is patte, a journey’s leg étape, a chair’s leg pied, a human leg jambe; foot is pied for humans but patte for animals.

   English	context	French
   leg	human	jambe
   leg	animal	patte
   leg	journey	étape
   leg	chair	pied
   foot	human	pied
   foot, paw	animal	patte

   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 280
10. What is a lexical gap? Give an example.
    When no word or phrase short of an explanatory footnote expresses a word’s exact meaning in another language: English has no neat equivalent of Mandarin xiào / Japanese oyakōkō (“filial piety”).
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 280
11. What is the difference between verb-framed and satellite-framed languages?
    Verb-framed languages mark the direction of motion on the verb and manner on satellites (Spanish La botella salió flotando “exited floating”). Satellite-framed languages mark direction on particles or phrases and manner on the verb (English The bottle floated out).
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 280
12. Which languages are verb-framed and which satellite-framed?
    Verb-framed: Japanese, Tamil, Romance, Semitic and Mayan languages. Satellite-framed: Chinese and non-Romance Indo-European languages like English, Swedish, Russian, Hindi, Farsi.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 280
13. What are pro-drop languages?
    Languages that can omit pronouns when the referent is given in the discourse, e.g. Spanish “[El jefe] dio con un libro. Ø Mostró su hallazgo…”.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, pp. 280–281
14. What is referential density, and what are “hot” and “cold” languages?
    How much a language uses explicit pronouns rather than zeros. Referentially sparse languages like Chinese or Japanese, which make the hearer infer more, are cold; more explicit languages are hot (after McLuhan’s hot vs cold media).
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 281
15. Why is translating from Chinese or Japanese into English hard with respect to pronouns?
    The model must identify each omitted argument (zero) and recover its referent in order to insert the right English pronoun.
    • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 13, §13.1, p. 281