Stephen Krashen

第八屆全国中小学英语阅读教学学术研讨会专题

培养儿童和青少年的阅读素养对培养学生核心素养、落实立德树人根本任务具有重要意义。2023年1月印发的《关于健全学校家庭社会协同育人机制的意见》明确了学校、家庭和社会在协同育人机制下的责任，并给出了具体建议。在新的教育形势下，家庭和学校作为对儿童和青少年实施教育的两个主阵地，必须协同合作，共同实现育人目标。但是在实际的阅读教育中，学校和家庭尚未找到协同提升学生阅读素养的有效途径。本期“第八届全国中小学英语阅读教学学术研讨会专题”特别邀请美国南加利福尼亚大学荣休教授、著名语言教育家斯蒂芬·克拉申（Stephen Krashen）深入阐释了有趣和可理解性阅读的重要性。此外，专题内其他两篇文章分别从理论建设和实践的角度深入剖析了学校和家庭如何共同协作推进家校协同阅读育人。希望广大教师进一步学习和思考，不断探索更加科学、有效的家校协同阅读育人方式，促进学生阅读素养的全面发展。

Abstract： Most people would agree that reading is helpful in acquiring language. I present here a much stronger position： Genuinely interesting and comprehensible reading is not only the main source of literacy and language， but also an important source of knowledge in a wide number of areas， and also contributes to our understanding of others， otherwise known as empathy.

Keywords： pleasure reading; grammar; spelling; knowledge; empathy

Reading and language / literacy development

A substantial amount of research shows that students who participate in in-school reading programs （sustained silent reading， self-selected and other forms of extensive reading） do as well or better on tests of reading than comparisons in traditional programs （reviewed in Krashen， 2004）.

In the first part of this paper， I focus specifically on time spent “just reading” versus time spent on the study of grammar. I then present evidence that pleasure reading leads to more knowledge in a variety of topics and greater understanding of others （empathy）.

An important tool for doing this is multiple regression， which allows us to pit competing hypotheses against each other in a single analysis. I illustrate the method by presenting data from Stokes et al. （1998）， an analysis of predictors of the ability to use the Spanish subjunctive in a “monitor-free” situation， a communicative situation in which subjects， English speakers who had studied Spanish， were not focused on form. （Subjects were asked to complete sentences such as “Siempre studio quando…”.） Only data from subjects who said they were not aware that the test probed competence in the use of the subjunctive was included in the analysis.

In Table 1， the “beta” statistic presents the strength of each predictor， uninfluenced by the others. For example， it may be the case that two predictors of literacy are related to each other， e.g. “study” （years of formal study of Spanish） and years of “residence” in a Spanish-speaking country （those students who had more years of study of Spanish also lived in a Spanish-speaking country longer）. Multiple regression controls for this， and allows a comparison as if there were no relationship among the predictors.

The strongest predictor， by far， was “reading”， the time spent doing free voluntary pleasure reading done in Spanish. As presented in Table 1， “reading” was stronger than all other predictors including “subjunctive study”， the amount of formal study specifically of the subjunctive， and the amount of time living in the country where the language was spoken （“residence”）. “Reading” was the only predictor that was statistically significant.

Thus， comprehensible input in the form of self-selected reading defeated skill-building， as represented by “study” and “subjunctive study”.  The analysis also suggests that for late-acquired aspects of language such as the subjunctive， the everyday input one gets simply by living in the country where the language is spoken is not enough. You need to read.

Lee et al. （1996） reported nearly identical results. Their subjects were native speakers of Korean who were students of English as a foreign language while living in the US.  They were given two kinds of tests： （1） Grammaticality judgement of English sentences （e.g. Which is correct， “She is the woman who helped me”， or “She is the woman which helped me”？） and （2） A translation test （Korean to English）.

Predictors， as was the case in Stokes et al. （1998）， were years of formal study of English and length of residence in the US. As presented in Table 2， once again the only significant predictor was the amount of reading reported in English.

A common view is that everyday use of a language in interaction with native speakers is an effective way of improving competence in that language. Once again， multiple regression comes down on the side of reading.  Subjects in Gradman and Hanania （1991） were international students enrolled in intensive English classes at Indiana University who took a well-established test of academic English， the TOEFL. There was no measure of the amount of English study students did or of the length of residence in the US， but informal use of the language （“extracurricular speaking”） was slightly negatively related to TOEFL scores， as was “total exposure” to English.  In agreement with the studies reviewed so far， “extracurricular reading” was a significant predictor of TOEFL performance.

1. A case study

A single subject case study （Lin et al.， 2007） also provides a direct comparison of self-selected pleasure reading and traditional study. No statistical analysis was done， nor was it necessary. The two treatments compared here did not overlap in time， and the same measure was used to determine the effectiveness of study and reading.

“Sophia” immigrated to the US with her family from  Chinas Taiwan when she was in Grade 6. At that time， she had only minimal competence in English， the result of private lessons several days per week for two years.

Starting in Grade 8， Sophia was tested in English reading on the Idaho Standards Achievement Test （ISAT） each year in the fall and again in the spring. The expectation was that her scores would improve during the school year， because of formal English classes.

This is not what happened. During the academic year Sophias scores on the ISAT declined （！）， but when she returned to school in the fall， after the summer break， her scores were higher than they had been one year before： During the summer， she made substantial gains on the ISAT， making up for what she had lost during the academic year and then some. Clearly what Sophia had been doing during the summer was far more effective than the formal instruction she had during the academic year.

What did she do during the summer？ Pleasure reading， and all from books she selected herself. She escaped the heat of the summer in a local public library， and read books for pleasure， averaging about 50 books each summer. She did not read “classics” but read Nancy Drew， Sweet Valley High and books from the Christy Miller series. Her intention was not to improve her English but to enjoy a good book in a cool environment.

2. The PIRLS study

A study involving thousands of students in dozens of countries examining predictors of high scores on the Progress in International Reading Literacy Study （PIRLS） examination produced results remarkably similar to other research reviewed here （Lao et al.， 2021）.  I report in this section on three administrations of the PIRLS examination， all involving 10-year-olds， data collected in 2006 （45 countries and regions）， 2011 （57 countries and regions）， and 2016 （61 countries and regions）.

The PIRLS examination measures “reading for literary experience and reading to acquire and use information”. Students take the test in the national language of their country. Sample sizes in each administration ranged from 3349 to 18，245 students.

The 2006 study is the only one of this set that asked about “independent reading”， defined as the percentage of students who read independently in school every day or almost every day. More independent reading was related to higher PIRLS scores， but the results fell just short of statistical significance.

In all of three of the PIRLS studies （Table 3.1—Table 3.3）， （1） higher levels of poverty （SES： socioeconomic status） were related to lower performance on the PIRLS， （2） the presence of school library was related to higher scores， and （3） more instruction was related to lower scores of the PIRLS or had no effect. These results are consistent with those of the smaller scale studies discussed earlier. The presence of a classroom library had no significant effect， most likely due to the small selection available， nor did parental reading habits.

3. Spelling

Studies of spelling have not been done in a way similar to the reading-grammar studies described earlier， but there is interesting research showing impressive development of spelling without instruction， and the likely source was reading.

In Hammill et al. （1977）， students who had spelling instruction （using one of several different spelling series， see Table 4） clearly did better on spelling tests than uninstructed children in Grades 3 & 4. The advantage， however， disappeared in Grades 5 & 6. A plausible explanation is that by then children in Grades 5 & 6 had developed a reading habit.

Additional evidence that spelling can be improved without instruction comes from Curtiss and Dolch （1939）， who reported that children improved on words taught and tested in previous years： third， fourth and fifth graders scored 71.8%， 79.6% and 83.7% respectively on the Grade 3 spelling test. Students in each grade also did better on words that were never taught or tested at all （Grade 4， 23.1%; Grade 6，/ 65.5%， Grade 8， 82.3%）.  Ormrod （1986） also reported that readers who read texts that contained unusual names （e.g. Aerodern） learned to spell about 30% of the names from reading alone.

4. Libraries

Still more evidence that reading is the major source of literacy development comes from studies by Keith Lance， who provided research showing that literacy development improves among young readers when school libraries are well-supplied with books and have the services of credentialed librarians1.

Also， Lance and Marks （2008） reported a strong positive link between circulation of childrens materials in local public libraries and 4th grade reading scores.

Unless it can be demonstrated that libraries function mainly as study halls and recreational centers， these results support the “reading hypothesis” that self-selected reading is responsible for the development of literacy.

Reading and knowledge

It is likely that many people understand that reading has at least some value in increasing literacy， but nearly unknown is the consistent finding that those who read more， including reading fiction， know more.

The major study in this area is Stanovich and Cunningham （1993）. Freshmen and sophomore college students took tests on a variety of topics， including science， social studies， current events， personal finance， health， “daily living technology”， cultural knowledge， and “multicultural literacy”.  （An example of a finance question was “What is the term for the amount of money charged for a loan and calculated as a percentage of that loan？”; a science question was “In what part of the body does the infection called pneumonia occur？”; a social science question was “Where is the Panama Canal？”.） Subjects were also asked if they recognized names such as Linus Pauling， Isaac Newton， and Bertrand Russell. In short， the tests as a group included the general knowledge we expect high school graduates to have.

Subjects were also asked about their familiarity with current authors and magazines， using authors such as Stephen King and Maya Angelou， and magazines such as Forbes and Ladies Home Journal. These two measures combined to make up a measure of “print exposure”.

Print exposure was by far the best predictor of the combined test battery described above. Of great interest， high school grades were a much weaker predictor of performance on tests of knowledge， and did not reach statistical significance. Performance on tests of reading comprehension， mathematics and analytic thinking （Ravens Progressive Matrices） reached statistical significance but were weaker predictors than print exposure. Exposure to TV had no value as a predictor of general knowledge （see Table 5）. Once again， it is reading that counts—not even the hard study and homework that students do that results in higher grades seems to improve general knowledge.

Reading and understanding others

Fiction readers develop more understanding of others. As Terry Gross of National Public Radios Fresh Air put it， when you read fiction “Youre learning to be somebody else， learning to see the world through their eyes.” 2

Fiction readers also appreciate the complexity of life and avoid overly simple solutions to problems. Experimental evidence for these conclusions can be found at Kidd & Castano （2013） and Djikic et al. （2013）.

Others have arrived at similar views， people who have most likely not read the scholarly research：

When I think about how I understand my role as citizen… the most important stuff Ive learned， I think Ive learned from novels. It has to do with empathy. It has to do with being comfortable with the notion that the world is complicated and full of grays， but theres still truth there to be found ... its possible to connect with some（one） else even though theyre very different from you. （Obama， 2015） 3

Chomsky （1988） also appreciated the value of fiction： “It is quite possible—overwhelmingly probable… that we will always learn more about human life and personality from novels than from scientific psychology.” （p. 159）

Some conclusions

Reading for pleasure was once disdained. Bean （1879） warned us that “… the craze for books leads to inattention， want of application， distaste for study， and unretentive memory.” （p. 347） An editorial in the Christian Guardian （7/31/1850） told readers that “No dissipation can be worse than that induced by the perusal of exciting books of fiction.” （Graff， 1979， p. 39） （For a more complete discussion of the rejection of fiction， see Nell， 1988）. But since then， evidence has accumulated and shown that fiction builds not only literacy， but also knowledge of a number of areas and a better understanding of other people.

The response to the pressure to raise achievement has been to increase homework， but it is possible that the time spent on homework would be better spent doing other things that give young people more time for reading.

Kohn （2007） summarizes the research on homework as follows： ... there is absolutely no evidence of any academic benefit from assigning homework in elementary or middle school. For younger students， in fact， there isnt even a correlation between whether children do homework （or how much they do） and any meaningful measures of achievement. At the high school level， the correlation is weak and tends to disappear when more sophisticated statistical measures are applied.

Kohn （2006） endorses replacing homework with self-selected reading： Some of the most thoughtful elementary school teachers I meet tell me that the only homework they give is to ask children to read books of their own choosing. This is a satisfying policy， first， because sustained reading... helps children to become more proficient readers. In fact， the research supporting that conclusion is as powerful as the research supporting homework is weak.

Even modest decreases in homework and modest improvements in providing access to interesting reading material might make a substantial difference in achievement， as well as making life more interesting and pleasant for teachers and students.

Implications

The results of these studies are spectacular. In every case， instruction is the loser. In all cases where it was part of the analysis， self-selected reading does very well.

It is not surprising that these results are not better known， as they are buried in professional journals and books， which are often challenging for non-specialists to read.  They are also expensive—so expensive that unless one has access to a well-supported professional university library， their cost makes it impossible to stay current. （Some professional journals will provide readers with copies of individual papers， but the usual cost is about $40 per paper. The money goes to the journal and the publisher， not the author.） Professional journal subscriptions cost hundreds of dollars per year， and scholars need to keep track of dozens of journals to keep in contact with the research.

At the same time， research supporting the role of public and school libraries has increased and consistently shows that students who attend schools with quality libraries （collections） and the services of a professional school librarian show better results on measures of reading.

School and public libraries， however， are now under unprecedented attack. A clear sign of this is the book banning movement， which seems to be limited to libraries， with little interest in banning the sales of certain books in bookstores. Not only are books banned， but libraries in several states are in danger of being defunded because of the availability of some books which have been classified as “harmful” to young readers.

I suspect that defunding and book banning are， at least in part， an attack on libraries， an attack on being able to read books for free. Publishers do not make much money when a school or public library orders one copy of a book. But if a book is banned， it is free publicity and interest in reading （and buying） the book increases. Mark Twains response when The Adventures of Huckleberry Finn was banned was immediate. He told his publisher： “That will sell 25，000 copies for us， sure.”

References

Bean， M. 1879. The evil of unlimited freedom in the use of juvenile fiction [J]. Library Journal， （4）： 341—343

Chomsky， N. 1988. Language and problems of knowledge [M]. Cambridge， MA： The MIT Press.

Curtiss， H. & Dolch， E. 1939.  Do spelling books teach spelling？ [J]. Elementary School Journal， 39（8）： 584—592

Djikic， M.， Oatley， K. & Moldoveanu， M. C. 2013. Opening the closed mind： The effect of exposure to literature on the need for closure [J]. Creativity Research Journal， 25（2）： 149—154

Gradman， H. L. & Hanania， E. 1991. Language learning background factors and ESL proficiency [J]. Modern Language Journal， 75（1）： 39—51

Graff， H. 1979. The literacy myth： Literacy and social structure in the nineteenth-century city [M].  New York， NY： Academic Press.

Hammill， D.， Larsen， S. & McNutt， G. 1977.  The effect of spelling instruction： A preliminary study [J]. Elementary School Journal， 78（1）： 67—72

Kidd， D. & Castano， E. 2013. Reading literary fiction improves theory of mind [J]. Science， 342 （6156）： 377—380

Kohn， A. 2006. The homework myth： Why our kids get too much of a bad thing [M]. Burlington， VT： Da Capo Books.

Kohn， A. 2007. Rethinking homework [J]. Principal， 86（3）： 35—38

Krashen， S. 1994. The input hypothesis and its rivals [A]. In N. Ellis （ed.）. Implicit and Explicit Learning of Languages [C]. London： Academic Press： 45—77

Krashen， S. 2004. The power of reading [M]. Santa Barbara， CA： Libraries Unlimited.

Krashen， S. 2011. Free voluntary reading [M]. Santa Barbara， CA： Libraries Unlimited.

Lance， K. C. & Marks， R. B. 2008. The link between public libraries and early reading success [J]. School Library Journal， 54（9）： 44—47

Lao， C.， Lee， S-Y.， McQuillan， J. & Krashen， S. 2021. Predicting reading ability among ten-year olds： Poverty （negative）， school libraries （positive）， instruction （zero）， early literacy （zero） [J]. Language Magazine， 20（10）： 20—21

Lee， S. Y. 2005. Facilitating and inhibiting factors on EFL writing： A model testing with SEM [J]. Language Learning， 55（2）： 335—374

Lee， Y. O.， Krashen， S. & Gribbons， B. 1996. The effect of reading on the acquisition of English relative clauses [J]. ITL： Review of Applied Linguistics， 113（1）： 263—273

Lin， S-Y， Shin， F. & Krashen， S.  2007. Sophias choice： Summer reading [J]. Knowledge Quest， 35（4）： 52—55

Nell， V. 1988. Lost in a book： The psychology of reading for pleasure [M]. New Haven， CT： Yale University Press.

Ormrod， J. E. 1986. Learning to spell： Three studies at the university level [J]. Research in the Teaching of English， 20（2）： 160—173

Stanovich， K. E.  & Cunningham， A. E.  1993. Where does knowledge come from？ Specific associations between print exposure and information acquisition [J]. Journal of Educational Psychology， 85（2）： 211—229

Stokes， J.， Krashen， S. & Kartchner， J. 1998. Factors in the acquisition of the present subjunctive in Spanish： The role of reading and study [J]. ITL： Review of Applied Linguistics， 121（1）： 19—25