positional variation of phrasepositional variation of phrase...

22
Positional variation of phrase Positional variation of phrase Positional variation of phrase Positional variation of phrase- - - frames in a new corpus of frames in a new corpus of frames in a new corpus of frames in a new corpus of proficient student writing proficient student writing proficient student writing proficient student writing Ute Römer & Matthew Brook O‘Donnell AACL Edmonton, Canada – 9 October 2009 www.elicorpora.info www.elicorpora.info www.elicorpora.info www.elicorpora.info

Upload: others

Post on 22-Jan-2021

10 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

Positional variation of phrasePositional variation of phrasePositional variation of phrasePositional variation of phrase----frames in a new corpus of frames in a new corpus of frames in a new corpus of frames in a new corpus of proficient student writingproficient student writingproficient student writingproficient student writingproficient student writingproficient student writingproficient student writingproficient student writing

Ute Römer & Matthew Brook O‘Donnell

AACL Edmonton, Canada – 9 October 2009

www.elicorpora.infowww.elicorpora.infowww.elicorpora.infowww.elicorpora.info

Page 2: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

Presentation outline

1.1.1.1. Introduction to MICUSP Introduction to MICUSP Introduction to MICUSP Introduction to MICUSP (Michigan Corpus of (Michigan Corpus of (Michigan Corpus of (Michigan Corpus of

UpperUpperUpperUpper----level Student Papers)level Student Papers)level Student Papers)level Student Papers): Composition and : Composition and : Composition and : Composition and markupmarkupmarkupmarkup

2.2.2.2. Positional variation of phrasePositional variation of phrasePositional variation of phrasePositional variation of phrase----frames in frames in frames in frames in MICUSPMICUSPMICUSPMICUSPMICUSPMICUSPMICUSPMICUSP

2.1 Method2.1 Method2.1 Method2.1 Method

2.2 Select results2.2 Select results2.2 Select results2.2 Select results

3.3.3.3. ConclusionConclusionConclusionConclusion

Ute RUte RUte RUte Röööömermermermer ([email protected])Matthew Brook O’Donnell Matthew Brook O’Donnell Matthew Brook O’Donnell Matthew Brook O’Donnell ([email protected])

Page 3: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

MICUSP backgroundMICUSP backgroundMICUSP backgroundMICUSP background

� Rather extensive research on academic writing Rather extensive research on academic writing Rather extensive research on academic writing Rather extensive research on academic writing produced by experts and by learners (CL and EAP)produced by experts and by learners (CL and EAP)produced by experts and by learners (CL and EAP)produced by experts and by learners (CL and EAP)

� Gap: advanced but unpublished academic writing Gap: advanced but unpublished academic writing Gap: advanced but unpublished academic writing Gap: advanced but unpublished academic writing by graduateby graduateby graduateby graduate----level university studentslevel university studentslevel university studentslevel university students

1. Introduction to MICUSP: Composition and markup

� Why? Difficulty of accessing unpublished academic Why? Difficulty of accessing unpublished academic Why? Difficulty of accessing unpublished academic Why? Difficulty of accessing unpublished academic writingwritingwritingwriting--------especially in any systematic wayespecially in any systematic wayespecially in any systematic wayespecially in any systematic way

"Far more academic writing is produced for assessment "Far more academic writing is produced for assessment "Far more academic writing is produced for assessment "Far more academic writing is produced for assessment purposes than for publication purposes, but because of purposes than for publication purposes, but because of purposes than for publication purposes, but because of purposes than for publication purposes, but because of the the the the lack of a suitable corpuslack of a suitable corpuslack of a suitable corpuslack of a suitable corpus, research into the generic , research into the generic , research into the generic , research into the generic features of published academic writing vastly outweighs features of published academic writing vastly outweighs features of published academic writing vastly outweighs features of published academic writing vastly outweighs research into the generic features of assessed student research into the generic features of assessed student research into the generic features of assessed student research into the generic features of assessed student writing" (writing" (writing" (writing" (NesiNesiNesiNesi et al. 2004: 440)et al. 2004: 440)et al. 2004: 440)et al. 2004: 440)

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 4: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

� Over 800 AOver 800 AOver 800 AOver 800 A----graded papers; around graded papers; around graded papers; around graded papers; around 2.3 million 2.3 million 2.3 million 2.3 million words words words words (May 2009 pre(May 2009 pre(May 2009 pre(May 2009 pre----release version)release version)release version)release version)

� Papers collected from Papers collected from Papers collected from Papers collected from 16 disciplines 16 disciplines 16 disciplines 16 disciplines across 4 across 4 across 4 across 4 academic divisions academic divisions academic divisions academic divisions (Humanities & Arts; Social Sciences; (Humanities & Arts; Social Sciences; (Humanities & Arts; Social Sciences; (Humanities & Arts; Social Sciences; Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)

1. Introduction to MICUSP: Composition and markup

Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)Biological & Health Sciences; Physical Sciences)

� Students at Students at Students at Students at 4 levels 4 levels 4 levels 4 levels of study of study of study of study (senior undergraduates; (senior undergraduates; (senior undergraduates; (senior undergraduates; 1st, 2nd, 3rd year graduates)1st, 2nd, 3rd year graduates)1st, 2nd, 3rd year graduates)1st, 2nd, 3rd year graduates)

� Native and nonNative and nonNative and nonNative and non----native speaker contributionsnative speaker contributionsnative speaker contributionsnative speaker contributions

� Freely accessible online by the end of 2009Freely accessible online by the end of 2009Freely accessible online by the end of 2009Freely accessible online by the end of 2009

� More info: flyer and http://micusp.elicorpora.infoMore info: flyer and http://micusp.elicorpora.infoMore info: flyer and http://micusp.elicorpora.infoMore info: flyer and http://micusp.elicorpora.info

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 5: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

Distribution of papers across disciplinesDistribution of papers across disciplinesDistribution of papers across disciplinesDistribution of papers across disciplines

1. Introduction to MICUSP: Composition and markup

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 6: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

MICUSP MICUSP MICUSP MICUSP markupmarkupmarkupmarkup

� Each paper is encoded in TEIEach paper is encoded in TEIEach paper is encoded in TEIEach paper is encoded in TEI----compliant XMLcompliant XMLcompliant XMLcompliant XML

� Combination of automatic and manual codingCombination of automatic and manual codingCombination of automatic and manual codingCombination of automatic and manual coding

� File header incorporating the metadata collected File header incorporating the metadata collected File header incorporating the metadata collected File header incorporating the metadata collected during paper submissionduring paper submissionduring paper submissionduring paper submission

1. Introduction to MICUSP: Composition and markup

during paper submissionduring paper submissionduring paper submissionduring paper submission

� Structural divisions Structural divisions Structural divisions Structural divisions (headings, sections, (headings, sections, (headings, sections, (headings, sections, paragraphs) of the original paper maintainedparagraphs) of the original paper maintainedparagraphs) of the original paper maintainedparagraphs) of the original paper maintained

� Sentence tokenizationSentence tokenizationSentence tokenizationSentence tokenization

� Encoding of textual features like quotations, Encoding of textual features like quotations, Encoding of textual features like quotations, Encoding of textual features like quotations, emphasis, bulletsemphasis, bulletsemphasis, bulletsemphasis, bullets

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 7: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2. Positional variation of phrase-frames in MICUSP

2.1 Method2.1 Method2.1 Method2.1 Method

� Extracted nExtracted nExtracted nExtracted n----grams and pgrams and pgrams and pgrams and p----frames (spans: 3, 4, 5; frames (spans: 3, 4, 5; frames (spans: 3, 4, 5; frames (spans: 3, 4, 5; floor=1) using floor=1) using floor=1) using floor=1) using kfNgramkfNgramkfNgramkfNgram ((((Fletcher 2002Fletcher 2002Fletcher 2002Fletcher 2002----2007)2007)2007)2007)

� PPPP----frames reduce nframes reduce nframes reduce nframes reduce n----gram lists in a motivated waygram lists in a motivated waygram lists in a motivated waygram lists in a motivated way� remove topicremove topicremove topicremove topic----specific items while highlighting specific items while highlighting specific items while highlighting specific items while highlighting

discourse itemsdiscourse itemsdiscourse itemsdiscourse items

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

discourse itemsdiscourse itemsdiscourse itemsdiscourse items� are more suited for the study of intraare more suited for the study of intraare more suited for the study of intraare more suited for the study of intra----textual textual textual textual

variationvariationvariationvariation

� Restricted definition of pRestricted definition of pRestricted definition of pRestricted definition of p----frames used here: only frames used here: only frames used here: only frames used here: only items with an internal variable slot items with an internal variable slot items with an internal variable slot items with an internal variable slot ((((RömerRömerRömerRömer, , , , forthcforthcforthcforthc.).).).), , , , not *BCD/ABC* type not *BCD/ABC* type not *BCD/ABC* type not *BCD/ABC* type ((((BiberBiberBiberBiber 2009)2009)2009)2009)

� Different degrees of variability (VPR)Different degrees of variability (VPR)Different degrees of variability (VPR)Different degrees of variability (VPR)

Page 8: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2. Positional variation of phrase-frames in MICUSP

2.1 Method 2.1 Method 2.1 Method 2.1 Method –––– HighHighHighHigh----frequency pfrequency pfrequency pfrequency p----framesframesframesframes

Span 3Span 3Span 3Span 3 Span 4Span 4Span 4Span 4 Span 5Span 5Span 5Span 5

the * ofthe * ofthe * ofthe * ofto * theto * theto * theto * thethe * andthe * andthe * andthe * anda * ofa * ofa * ofa * of

the * of thethe * of thethe * of thethe * of thein the * ofin the * ofin the * ofin the * ofof the * ofof the * ofof the * ofof the * ofto the * ofto the * ofto the * ofto the * of

in the * of thein the * of thein the * of thein the * of theof the * of theof the * of theof the * of theof the * of theto the * of theto the * of theto the * of theto the * of theon the * of theon the * of theon the * of theon the * of the

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

a * ofa * ofa * ofa * ofof * andof * andof * andof * andthe * tothe * tothe * tothe * tothe * thatthe * thatthe * thatthe * thatthe * inthe * inthe * inthe * inand * theand * theand * theand * thethe * isthe * isthe * isthe * is

to the * ofto the * ofto the * ofto the * ofand the * ofand the * ofand the * ofand the * ofthe * of athe * of athe * of athe * of aon the * ofon the * ofon the * ofon the * ofof the * andof the * andof the * andof the * andfor the * offor the * offor the * offor the * ofthe * and thethe * and thethe * and thethe * and the

on the * of theon the * of theon the * of theon the * of theat the * of theat the * of theat the * of theat the * of theand the * of theand the * of theand the * of theand the * of thein order to * thein order to * thein order to * thein order to * thewith the * and thewith the * and thewith the * and thewith the * and theof the * and theof the * and theof the * and theof the * and thefor the * of thefor the * of thefor the * of thefor the * of the

Page 9: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2. Positional variation of phrase-frames in MICUSP

2.1 Method 2.1 Method 2.1 Method 2.1 Method ---- Recording textual position Recording textual position Recording textual position Recording textual position

� “textual colligation” “textual colligation” “textual colligation” “textual colligation” ((((HoeyHoeyHoeyHoey 2005)2005)2005)2005)

� XML XML XML XML markupmarkupmarkupmarkup of document structure allows at of document structure allows at of document structure allows at of document structure allows at least three kinds of position to be recorded:least three kinds of position to be recorded:least three kinds of position to be recorded:least three kinds of position to be recorded:

1. Position of 11. Position of 11. Position of 11. Position of 1stststst word of the item word of the item word of the item word of the item within its sentencewithin its sentencewithin its sentencewithin its sentence

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

1. Position of 11. Position of 11. Position of 11. Position of 1stststst word of the item word of the item word of the item word of the item within its sentencewithin its sentencewithin its sentencewithin its sentence

2. Position of item’s sentence 2. Position of item’s sentence 2. Position of item’s sentence 2. Position of item’s sentence within the paragraphwithin the paragraphwithin the paragraphwithin the paragraph

3. Position of the paragraph containing the item 3. Position of the paragraph containing the item 3. Position of the paragraph containing the item 3. Position of the paragraph containing the item within within within within the textthe textthe textthe text

� Creation of a pCreation of a pCreation of a pCreation of a p----frame/nframe/nframe/nframe/n----gram and positional gram and positional gram and positional gram and positional variation database from which distributions variation database from which distributions variation database from which distributions variation database from which distributions (for (for (for (for

positions 1, 2, 3)positions 1, 2, 3)positions 1, 2, 3)positions 1, 2, 3) can be retrievedcan be retrievedcan be retrievedcan be retrieved

Page 10: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.1 Method 2.1 Method 2.1 Method 2.1 Method ---- Position Position Position Position of of of of phrase within sentencephrase within sentencephrase within sentencephrase within sentence

� For each instance of For each instance of For each instance of For each instance of pppp----frame in frame in frame in frame in MICUSP, identify MICUSP, identify MICUSP, identify MICUSP, identify position of first word position of first word position of first word position of first word within its sentencewithin its sentencewithin its sentencewithin its sentence

� Divide sentence into 3 Divide sentence into 3 Divide sentence into 3 Divide sentence into 3

Pos 1 Pos 1 Pos 1 Pos 1 25%25%25%25%

Pos 2 Pos 2 Pos 2 Pos 2 50%50%50%50%

Pos 3 Pos 3 Pos 3 Pos 3 25%25%25%25%

2. Positional variation of phrase-frames in MICUSP

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

� Divide sentence into 3 Divide sentence into 3 Divide sentence into 3 Divide sentence into 3 parts: first and last 25%, parts: first and last 25%, parts: first and last 25%, parts: first and last 25%, remaining middle section remaining middle section remaining middle section remaining middle section (50%)(50%)(50%)(50%)

Pos 1Pos 1Pos 1Pos 1 Pos 2Pos 2Pos 2Pos 2 Pos 3Pos 3Pos 3Pos 3

5 4 2

Page 11: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.1 Method 2.1 Method 2.1 Method 2.1 Method ---- Position of sentence within paragraphPosition of sentence within paragraphPosition of sentence within paragraphPosition of sentence within paragraph

� For each instance of For each instance of For each instance of For each instance of pppp----frame in frame in frame in frame in MICUSP, identify MICUSP, identify MICUSP, identify MICUSP, identify location of containing location of containing location of containing location of containing sentence within its sentence within its sentence within its sentence within its paragraphparagraphparagraphparagraph

sentences

2. Positional variation of phrase-frames in MICUSP

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

paragraphparagraphparagraphparagraph

� Divide paragraph into 3 Divide paragraph into 3 Divide paragraph into 3 Divide paragraph into 3 parts: first (Pos 1) and last parts: first (Pos 1) and last parts: first (Pos 1) and last parts: first (Pos 1) and last sentence of paragraph sentence of paragraph sentence of paragraph sentence of paragraph (Pos 3), sentences that are (Pos 3), sentences that are (Pos 3), sentences that are (Pos 3), sentences that are not first or last (Pos 2)not first or last (Pos 2)not first or last (Pos 2)not first or last (Pos 2)

sentences

paragraphs

Page 12: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.1 Method 2.1 Method 2.1 Method 2.1 Method ---- Position of paragraph within textPosition of paragraph within textPosition of paragraph within textPosition of paragraph within text

� For each instance of For each instance of For each instance of For each instance of pppp----frame in frame in frame in frame in MICUSP, identify MICUSP, identify MICUSP, identify MICUSP, identify location of containing location of containing location of containing location of containing paragraph within its paragraph within its paragraph within its paragraph within its text/papertext/papertext/papertext/paper

15%15%15%15%

2. Positional variation of phrase-frames in MICUSP

text/papertext/papertext/papertext/paper

� Divide paragraph into 3 Divide paragraph into 3 Divide paragraph into 3 Divide paragraph into 3 parts: parts: parts: parts: Paragraph is within first Paragraph is within first Paragraph is within first Paragraph is within first 15% of paper (Pos 1), paragraph 15% of paper (Pos 1), paragraph 15% of paper (Pos 1), paragraph 15% of paper (Pos 1), paragraph is part of midis part of midis part of midis part of mid----70% of paper 70% of paper 70% of paper 70% of paper (Pos 2), paragraph is within (Pos 2), paragraph is within (Pos 2), paragraph is within (Pos 2), paragraph is within final 15% of paper (Pos 3)final 15% of paper (Pos 3)final 15% of paper (Pos 3)final 15% of paper (Pos 3)

Pos 1Pos 1Pos 1Pos 1 Pos 2Pos 2Pos 2Pos 2 Pos 3Pos 3Pos 3Pos 3

1 1 4

paragraphs

15%15%15%15%

70%70%70%70%

Page 13: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: the * thatthe * thatthe * thatthe * that -Avoidance of Avoidance of Avoidance of Avoidance of sentencesentencesentencesentence----final final final final positionpositionpositionposition

-Even distribution Even distribution Even distribution Even distribution across paragraphsacross paragraphsacross paragraphsacross paragraphs

-Slight textSlight textSlight textSlight text----final final final final

2. Positional variation of phrase-frames in MICUSP

VPR: VPR: VPR: VPR: 33.49%33.49%33.49%33.49%Variant/PVariant/PVariant/PVariant/P----frame Ratio (Rframe Ratio (Rframe Ratio (Rframe Ratio (Röööömermermermer, , , , forthcforthcforthcforthc.).).).)

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

-Slight textSlight textSlight textSlight text----final final final final preferencepreferencepreferencepreference

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

���� ����

Page 14: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: it is * thatit is * thatit is * thatit is * that

2. Positional variation of phrase-frames in MICUSP

-Strong sentenceStrong sentenceStrong sentenceStrong sentence----initial preferenceinitial preferenceinitial preferenceinitial preference

-Even distribution Even distribution Even distribution Even distribution across paragraphsacross paragraphsacross paragraphsacross paragraphs

-Preference for Preference for Preference for Preference for texttexttexttext----final positionfinal positionfinal positionfinal position

VPR: VPR: VPR: VPR: 19.05%19.05%19.05%19.05%

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

texttexttexttext----final positionfinal positionfinal positionfinal position

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

���� ����

Page 15: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: it would be * toit would be * toit would be * toit would be * to

----Strong preference Strong preference Strong preference Strong preference for sentencefor sentencefor sentencefor sentence----initial, initial, initial, initial, and textand textand textand text----final final final final positions positions positions positions

2. Positional variation of phrase-frames in MICUSP

VPR: VPR: VPR: VPR: 37.25%37.25%37.25%37.25%

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

���� ����

Page 16: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: as * in figureas * in figureas * in figureas * in figure

2. Positional variation of phrase-frames in MICUSP

----Strong Strong Strong Strong preference for preference for preference for preference for sentencesentencesentencesentence----final final final final position position position position

----Mild paragraphMild paragraphMild paragraphMild paragraph----and textand textand textand text----final final final final

VPR: VPR: VPR: VPR: 21.57%21.57%21.57%21.57%

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

and textand textand textand text----final final final final preference preference preference preference

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

����

Page 17: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: in order to * thein order to * thein order to * thein order to * the

2. Positional variation of phrase-frames in MICUSP

----Strong preference Strong preference Strong preference Strong preference for sentencefor sentencefor sentencefor sentence---- and and and and paragraphparagraphparagraphparagraph---- initial initial initial initial positions positions positions positions ----Mild textMild textMild textMild text----medial medial medial medial preferencepreferencepreferencepreference

VPR: VPR: VPR: VPR: 61.64%61.64%61.64%61.64%

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

preferencepreferencepreferencepreference----Avoids sentenceAvoids sentenceAvoids sentenceAvoids sentence----final positionfinal positionfinal positionfinal position

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

���� ����

Page 18: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

2.2 Select 2.2 Select 2.2 Select 2.2 Select results: results: results: results: are * likely toare * likely toare * likely toare * likely to

2. Positional variation of phrase-frames in MICUSP

----Avoids sentenceAvoids sentenceAvoids sentenceAvoids sentence----final and final and final and final and favorsfavorsfavorsfavorssentencesentencesentencesentence----medial/ medial/ medial/ medial/ initial positionsinitial positionsinitial positionsinitial positions----Strongly prefers Strongly prefers Strongly prefers Strongly prefers paragraphparagraphparagraphparagraph----medial/ medial/ medial/ medial/ final positionsfinal positionsfinal positionsfinal positions----Mild textMild textMild textMild text----medial medial medial medial

VPR: VPR: VPR: VPR: 7.63%7.63%7.63%7.63%

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

----Mild textMild textMild textMild text----medial medial medial medial preferencepreferencepreferencepreference

χχχχ2222

p<0.001p<0.001p<0.001p<0.001

���� ����

Page 19: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

3. Conclusion

� We have introduced We have introduced We have introduced We have introduced MICUSP as a new resource MICUSP as a new resource MICUSP as a new resource MICUSP as a new resource for for for for the study of proficient student academic writingthe study of proficient student academic writingthe study of proficient student academic writingthe study of proficient student academic writing

� Use of pUse of pUse of pUse of p----frames as effective way to generalize frames as effective way to generalize frames as effective way to generalize frames as effective way to generalize phraseologicalphraseologicalphraseologicalphraseological items in in advanced student writingitems in in advanced student writingitems in in advanced student writingitems in in advanced student writing

� Explored method to investigate textual distribution Explored method to investigate textual distribution Explored method to investigate textual distribution Explored method to investigate textual distribution of pof pof pof p----framesframesframesframesof pof pof pof p----framesframesframesframes

Pedagogical implicationsPedagogical implicationsPedagogical implicationsPedagogical implications

� Important for novice writers to identify commonly Important for novice writers to identify commonly Important for novice writers to identify commonly Important for novice writers to identify commonly used phrases in discourse of a disciplineused phrases in discourse of a disciplineused phrases in discourse of a disciplineused phrases in discourse of a discipline

� Important for EAP teachers and novice academic Important for EAP teachers and novice academic Important for EAP teachers and novice academic Important for EAP teachers and novice academic writers to know which items/phrases to use where writers to know which items/phrases to use where writers to know which items/phrases to use where writers to know which items/phrases to use where in a textin a textin a textin a text

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 20: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

3. Conclusion

Future avenuesFuture avenuesFuture avenuesFuture avenues

� Look at the textual distribution of a larger number Look at the textual distribution of a larger number Look at the textual distribution of a larger number Look at the textual distribution of a larger number of items and use more rigorous statistical methodsof items and use more rigorous statistical methodsof items and use more rigorous statistical methodsof items and use more rigorous statistical methods

� Analyze pAnalyze pAnalyze pAnalyze p----frame internal variation frame internal variation frame internal variation frame internal variation (semantic (semantic (semantic (semantic grouping of variants)grouping of variants)grouping of variants)grouping of variants)

� Study frequent items in context to examine their Study frequent items in context to examine their Study frequent items in context to examine their Study frequent items in context to examine their � Study frequent items in context to examine their Study frequent items in context to examine their Study frequent items in context to examine their Study frequent items in context to examine their discourse functionsdiscourse functionsdiscourse functionsdiscourse functions

� Carry out comparisons with expert/published Carry out comparisons with expert/published Carry out comparisons with expert/published Carry out comparisons with expert/published writing writing writing writing (research articles from different disciplines) (research articles from different disciplines) (research articles from different disciplines) (research articles from different disciplines) and and and and with comparable sets of learner academic writing with comparable sets of learner academic writing with comparable sets of learner academic writing with comparable sets of learner academic writing (availability issues)(availability issues)(availability issues)(availability issues)

MiMiMiMichigan chigan chigan chigan CCCCorpus of orpus of orpus of orpus of UUUUpperpperpperpper----level level level level SSSStudent tudent tudent tudent PPPPapersapersapersapers

Page 21: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

Thank you!Thank you!Thank you!Thank you!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!Watch out for MICUSP coming soon!

Ute Römer & Matthew Brook O‘Donnell

[email protected], [email protected]

www.elicorpora.infowww.elicorpora.infowww.elicorpora.infowww.elicorpora.info

Page 22: Positional variation of phrasePositional variation of phrase ...aacl2009/PDFs/RoemerODonnell...Positional variation of phrasePositional variation of phrase----frames in a new corpus

ReferencesÄdel, A. and U. Römer. (In preparation). Research on proficient academic writing across disciplines and levels: Introducing MICUSP. Biber, D. (2009). A corpus-driven approach to formulaic language in English: Multi-word patterns in speech and writing. International Journal of Corpus Linguistics 14(3): 275-311.Fletcher, W. H. (2002-2007). KfNgram. Annapolis, MD: USNA.Hoey, M. (2005). Lexical Priming: A New Theory of Words and Language. London: Routledge.Hoey, M. and M. B. O’Donnell (2008). ‘The Beginning of Something Important? Corpus Evidence on the Text Beginnings of Hard News Stories’. In: Lewandowska-Tomaszczyk, B. (ed.), Corpus Linguistics, Computer Tools, and Applications-State of the Art. PALC 2007. Bern: Peter Lang.Bern: Peter Lang.Nesi, H., G. Sharpling and L. Ganobcsik-Williams. (2004). Student papers across the curriculum: Designing and developing a corpus of British student writing. Computers and Composition 21: 439-450.O’Donnell, M. B. and U. Römer. (In preparation). From student hard drive to web corpus: The design, compilation, annotation and online distribution of MICUSP.Römer, U. (Forthcoming). Establishing the phraseological profile of a text type: The construction of meaning in academic book reviews. English Text Construction 3(1).Römer, U. and S. Wulff. (Forthcoming). Applying corpus methods to writing research: Explorations of MICUSP. Journal of Writing Research (http://www.jowr.org).Wulff, S. and U. Römer. (Forthcoming). Becoming a proficient academic writer: Shifting lexical preferences in the use of the progressive. Corpora.Wulff, S., U. Römer and J. M. Swales. (In preparation). Attended/unattended this in academic writing: Qualitative and quantitative perspectives. Corpus Linguistics and Linguistic Theory.