The role of constructions in understanding predictability measures and their correspondence to word duration
Studies of word predictability in context show that words in English tend to be shorter if they are predictable from the next word, and to a lesser extent, if they are predictable from the previous word. Some studies distinguish function and content words, but otherwise have not considered grammatic...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
De Gruyter
2024-08-01
|
| Series: | Cognitive Linguistics |
| Subjects: | |
| Online Access: | https://doi.org/10.1515/cog-2023-0077 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849225133226459136 |
|---|---|
| author | Bybee Joan Brown Earl Kjar |
| author_facet | Bybee Joan Brown Earl Kjar |
| author_sort | Bybee Joan |
| collection | DOAJ |
| description | Studies of word predictability in context show that words in English tend to be shorter if they are predictable from the next word, and to a lesser extent, if they are predictable from the previous word. Some studies distinguish function and content words, but otherwise have not considered grammatical factors, treating all two-word sequences as comparable. Because function words are highly frequent, words occurring with them have low predictability. Highest predictability occurs within bigrams with two content words. Using the Buckeye corpus, we show that content word bigrams from different constructions vary widely in predictability, with adjective–noun and noun–noun sequences (content words within a noun phrase) having the highest scores. It is known that in adjective–noun sequences, the vowel of the adjective is shorter than in other positions. We study noun–noun sequences within the noun phrase and show that the first noun is shorter than in other contexts. It follows that the shorter duration of the first word when it is predictable from the second in many cases is due to the noun phrase construction and not necessarily the regulation of duration corresponding to predictable versus unpredictable information. We conclude that predictability studies must consider the constructions words occur in. |
| format | Article |
| id | doaj-art-d1645842b3344921af54261e32e84e77 |
| institution | Kabale University |
| issn | 0936-5907 1613-3641 |
| language | English |
| publishDate | 2024-08-01 |
| publisher | De Gruyter |
| record_format | Article |
| series | Cognitive Linguistics |
| spelling | doaj-art-d1645842b3344921af54261e32e84e772025-08-25T06:10:17ZengDe GruyterCognitive Linguistics0936-59071613-36412024-08-0135337740610.1515/cog-2023-0077The role of constructions in understanding predictability measures and their correspondence to word durationBybee Joan0Brown Earl Kjar11104University of New Mexico, Albuquerque, NM, USABrigham Young University, Provo, USAStudies of word predictability in context show that words in English tend to be shorter if they are predictable from the next word, and to a lesser extent, if they are predictable from the previous word. Some studies distinguish function and content words, but otherwise have not considered grammatical factors, treating all two-word sequences as comparable. Because function words are highly frequent, words occurring with them have low predictability. Highest predictability occurs within bigrams with two content words. Using the Buckeye corpus, we show that content word bigrams from different constructions vary widely in predictability, with adjective–noun and noun–noun sequences (content words within a noun phrase) having the highest scores. It is known that in adjective–noun sequences, the vowel of the adjective is shorter than in other positions. We study noun–noun sequences within the noun phrase and show that the first noun is shorter than in other contexts. It follows that the shorter duration of the first word when it is predictable from the second in many cases is due to the noun phrase construction and not necessarily the regulation of duration corresponding to predictable versus unpredictable information. We conclude that predictability studies must consider the constructions words occur in.https://doi.org/10.1515/cog-2023-0077constructionspredictabilityinformativityphonetic reductionword durationnoun phrase construction |
| spellingShingle | Bybee Joan Brown Earl Kjar The role of constructions in understanding predictability measures and their correspondence to word duration Cognitive Linguistics constructions predictability informativity phonetic reduction word duration noun phrase construction |
| title | The role of constructions in understanding predictability measures and their correspondence to word duration |
| title_full | The role of constructions in understanding predictability measures and their correspondence to word duration |
| title_fullStr | The role of constructions in understanding predictability measures and their correspondence to word duration |
| title_full_unstemmed | The role of constructions in understanding predictability measures and their correspondence to word duration |
| title_short | The role of constructions in understanding predictability measures and their correspondence to word duration |
| title_sort | role of constructions in understanding predictability measures and their correspondence to word duration |
| topic | constructions predictability informativity phonetic reduction word duration noun phrase construction |
| url | https://doi.org/10.1515/cog-2023-0077 |
| work_keys_str_mv | AT bybeejoan theroleofconstructionsinunderstandingpredictabilitymeasuresandtheircorrespondencetowordduration AT brownearlkjar theroleofconstructionsinunderstandingpredictabilitymeasuresandtheircorrespondencetowordduration AT bybeejoan roleofconstructionsinunderstandingpredictabilitymeasuresandtheircorrespondencetowordduration AT brownearlkjar roleofconstructionsinunderstandingpredictabilitymeasuresandtheircorrespondencetowordduration |