Summary

Definite / indefinite articles (noun delimiters) and bare nouns are problematic for many learners and teachers of EFL/ESL. My research on this, from cognitive semantic perspectives, makes possible more intuitive and sensible approaches to teaching, explaining, and practicing these grammatical juggernauts. Several smaller studies have been published (Phase 1), based mainly on analysis of student responses to pragmatic questionnaires. Two major data sets await analysis for large-scale data analysis projects (Phase 2). This has led so far to a more comprehensive understanding of noun phrase patterns, and better teaching methods and materials.

• 
  Methods used: Pragmatic judgment questionnaires, statistical analysis to be conducted in SQL and R. Previous small-scale initial studies with logistic regression in SAS.
• 
  Results: Several published research papers; two major research papers to be submitted in later 2024; teaching materials with more intuitive explanations and coverage of noun phrse patterns; a forthcoming online grammar course will be created, to appear in 2025.

Details

Definite / indefinite articles (noun delimiters, as I call them) and bare nouns have been problematic for many learners and teachers of EFL/ESL (English as a second / foreign language). Few linguists actually understand how these noun phrase (NP) patterns work, and teachers generally do not know how to teach them. My research approaches this from cognitive semantic perspectives, whjich makes possible more intuitive and sensible approaches to teaching, explaining, and practicing these grammatical juggernauts.

I have conducted some smaller, initial studies with smaller data sets. This mainly consists of learner responses to pragmatic judgment questionnairs (filling in blanks with appropriate delimiters), which have been analyzed with loglinear / logistic regression. Also, qualitative survey data show that Korean students do not really understand these NP patterns or the rules they were taught; they find that the traditional textbook rules do not work, or they cannot apply them. My qualitative and quantitative analyses of NP patterns indicate that these do not follow traditional rules, but the use of delimiters or bare NPs follow a set of core and extended semantic patterns. This includes semantic extensions of definiteness for various specialized uses of ‘a’ and ‘the’, and for bare singular and plural NPs. NPs followed by post-modifers also warrant separate attention, as they often are specialized patterns.

Several smaller studies have been published (Phase 1), based mainly on analysis of student responses to pragmatic questionnaires, and a qualitative survey. Two major data sets await analysis for large-scale data analysis projects (Phase 2). This includes a full data set of learner and L1 responses to pragmatic judgment questionnaires, to be analyzed in SQL and R; and a large L1 corpus data set, to be analyzed in Python and/or R.

So far, this research has led so far to a more comprehensive understanding of noun phrase patterns, and better teaching methods and materials. The teaching methods and explanations developed so far explain NP patterns much more clearly for students, and explain NP patterns far better than traditional rules.

Publications

• Lee, K. (2023). Schematic semantic conceptualization of English definite noun phrases. Poster presented at 16th International Cognitive Linguistics Conference, August 2020, Düsseldorf.
• Lee, K. (2021). A study on the viability of a cognitive approach to teaching articles. Studies in English Education, 26(1), 29-55. https://dx.doi.org/10.22275/SEE.26.1.02
• Lee, K. (2021). Teaching articles: Addressing university students’ misconceptions. Journal of Learner-centered Curriculum and Instruction, 21(2), 1-25. https://dx.doi.org/10.22251/jlcci.2021.21.2.1
• Lee, K. (2020). Definite and indefinite articles with abstract nouns in L2. Modern English Education, 21(4), https://dx.doi.org/1-10. 10.18095/meeso.2020.21.4.1
• Lee, K. (2017). A “the” or the “a”? L2 learner problems and patterns. Korea TESOL Journal, 13(2), 25-48.

Sample SAS code

PROC IMPORT DATAFILE=CSV
		    OUT=det14
		    DBMS=CSV
		    REPLACE;
RUN;
*PROC PRINT DATA=det14; *RUN;

data det14; 
infile '/folders/myshortcuts/d/researchdata/detsurvey2014.all.csv'  truncover firstobs=2 dsd; *dlm=" " n=1300;    
input subj $ l1$ item $ wort $ age gender $ major $ enyrs ausland pt1time pt2time group $ nfall syn1 $ syn2 $ syn3 $
 synpost $ num $ antwort $ orig $ agr $ concrete semvar lfreq fam;

proc freq data=det14 order=data;  
 tables item*antwort/ nocol nocum nopercent; 
 tables gender*major; *tables wort; 
 tables orig * agr / chisq nocum ; *plots(only)=(freqplot mosaicplot);  
 tables antwort * orig / chisq nocum nocol; *plots(only)=(freqplot mosaicplot); 
 tables orig*num/ chisq nocum nopercent;
 tables ausland*l1/ nocol nocum nopercent; 
 tables l1*agr / chisq nocum  nocol; *plots(only)=(freqplot mosaicplot); 
 tables syn2*agr  / chisq nocum  nocol; *plots(only)=(freqplot mosaicplot); 
 tables synpost*agr  / chisq nocum  nocol; *plots(only)=(freqplot mosaicplot);  
 tables agr*ausland  / chisq nocum nocol; * plots(only)=(freqplot mosaicplot);  
 tables agr*gender / chisq nocum nocol; 
 tables nfall*agr / chisq nocum nopercent nocol; 
 tables num*agr / chisq nocum nopercent nocol; 
proc logistic data=det14 order=data ; 
 class antwort num l1 orig agr syn2 synpost;
 model antwort = orig l1 ausland num nfall syn2 lfreq  / l=clogit stb expb rsq lackfit expb corrb parmlabel cl aggregate ; 
 oddsratio orig;    
 *contrast 'syn2' syn2 1 1 0 -2 0;  contrast 'a z' orig 1 0 -1; 
 contrast 'the z' orig 0 1 -1; contrast 'a the' orig 1 -1 0; 
 contrast 'num' num 1 -1; 
 contrast 'Korean cf. Chinese' l1 -1 0 0 1; 
run;
 
proc logistic data=det14 order=data; 
 class l1 antwort num orig agr syn1 syn2 synpost ;
 model agr = l1 syn2 synpost nfall synpost lfreq  / l=logit stb expb rsq parmlabel cl ; 
 *oddsratio orig;    
 contrast 'syn2-a1-n' syn2 1 0 0 -1 0;  contrast 'syn2-a2-n' syn2 0 1 0 -1 0 ; contrast 'syn2-a4-n' syn2 0 0 1 -1 0 ;
 contrast 'L1' l1 1 0 0 -1; contrast 'L1-kr-others' l1 3 -1 -1 -1; 
run;
 
proc catmod data=det14 order=data;
 direct ausland lfreq concrete semvar lfreq fam; 
 model agr = l1 syn2 synpost nfall num  lfreq   ; 
 *contrast 'syn2' syn2 1 0 0 -1 0 ; *contrast 'syn2b' syn2 -1 .5 .5 0 0 ; 
 contrast 'syn2-a1-n' syn2 1 0 0 -1;  contrast 'syn2-a2-n' syn2 0 1 0 -1 ; contrast 'syn2-a4-n' syn2 0 0 1 -1 ;
 contrast 'L1'  l1 1 0 1 ;
run;