Writing-Assisting
AI Enables Learners to Think and Write with Their Passive Vocabulary and Develop
it into Active Vocabulary
AI is a tool that
one can use foolishly or wisely
AI is just a tool for humans.
Humans can misuse this new tool foolishly or utilize it wisely.
A foolish use of writing-assisting
AI (translation and rewriting apps) is to uncritically adopt the language output
of the AI (English in our argument). That immediate use can result in sentences
that deviate from the writer's intentions. It does not help their learning
English, either.
AI can facilitate
the transformation of learners' comprehension vocabulary into presentation
vocabulary
A smart way to use AI is to utilize
it to learn English through its use (or use English while learning it). Users
should recognize that AI as an assisting tool only produces an imperfect draft.
However, they should also realize that AI's output expands their possibilities
for thought and expression in a foreign language. AI can both be a practical
and learning tool.
In other words, writing-assisting
AI should be used to provide English learners with opportunities to think and
express using their passive vocabulary rather than their active vocabulary in
the target language. AI can facilitate the use of their passive language and
change it to their means for presenting their thoughts.
In general, the number of active words
that people use at their disposal to express their thoughts is much smaller
than the number of passive words they use to understand the expressions of
others. For instance, many Japanese people enjoy reading the works of Soseki
Natsume, but few can write as Soseki.
When it comes to foreign languages,
learners' active vocabulary is extremely limited. English classes in Japan
still do not provide enough training in English presentation, and therefore,
learners' resources for expression are highly restricted.
On the other hand, English language
learners preparing for university entrance exams are trained in reading
comprehension to a much higher level, and their comprehension vocabulary
increases in the process. The vocabulary acquired from rote memorization may
only produce an incomplete understanding of texts, though. However, in any
case, the passive lexicon of learners who study English for university entrance
exams outnumbers their active vocabulary.
Learners are aware of this vast
difference between the vocabulary for comprehension and that for expressions.
Probably, some learners are reluctant to engage in English presentation
activities because they can only express themselves at a humiliatingly low
level, far below the level of their intelligence. If that is the case, one of
the challenges for English education in Japan is to develop learners'
comprehension vocabulary into presentation vocabulary.
What I do in my classes
Suppose students have an assignment
to write in English -- A writing assignment requiring extended reflective
writing, not an impromptu speaking assignment. (My assumption is that AI is
adequate for writing instruction but not for teaching speaking). To make my
point clearer, I tentatively define a "writing assignment" as one
that demands a long piece of writing in which students have to express their
thoughts on a relatively complex topic accurately. To be specific, think of it
as the level of a university writing class.
When learners have to express their
thoughts in their foreign language (English), they are only allowed to think
and write with their limited active vocabulary of English from start to finish.
That process can be an exercise in utilizing the potentiality of their
restricted resources. However, the vocabulary limit narrows the range of topics
they can address. Even if they choose a challenging subject, the quality of
their thoughts and drafts will diminish significantly.
Here I would like to share my
teaching experiences. My university requires students to complete an English
academic essay of more than 1,000 words in the second semester of their first
year. One year, a student of mine said he wanted to write about the potential
dangers of Genetically Modified Food. As I listened, he presented his argument
for a few minutes, describing the risks in detail. I marveled at his expertise
and encouraged him to write a high-quality essay.
However, when I assigned my
students to write an outline in English in the next class, I found that that
student’s thesis statement (i.e., the sentence at the end of the introduction
that concretely declares the essay's claim) was only "GM Food is bad. "
When students think and write in their active vocabulary in English, the
products are often crude or, to be blunt, miserable. Knowing his high
intelligence in Japanese, I felt troubled by the grave gap between his
intellect and English.
In the following year, I decided to
introduce machine translation (MT) in the last five weeks of my second-semester
classes. In the first ten weeks, students write short passages in English
without the help of AI, producing numerous mistakes. My role as a teacher is to
share the policy that "the school is where learners can make mistakes
without worry and learn from them." I avoid negative evaluation of
mistakes and encourage students to learn specific English expressions from
errors that we share in class.
At the same time, students go
through brainstorming and outlining, gradually preparing a 3,000-word Japanese
essay (as a rough draft of their final 1,000-word English paper). As the
instructor, I offer feedback on the completed Japanese texts to ensure
reasonable quality. Classmates comment on the products mutually, too.
For the remainder of the semester,
students critically read and rewrite the English output from MT (post-editing.)
Students revise MT’s English after checking if it expresses their intentions
accurately, if the grammar (especially articles, singular/plural, pronouns, and
tense) is consistent, and if some stylistic improvement is necessary.
In one class, students' essay
topics changed noticeably after the introduction of MT. Topics included
intellectual ones such as "Time from the Viewpoint of the Special and General
Theory of Relativity," "Free-Will as a Fictional Concept,"
"The Data Revolution in American Baseball," "The Reasons Behind Improved
Performance in Track and Field," and "The Restraint of Emotional Expression
in Haiku."
Later, I conducted a questionnaire
in that class. In response to the prompt statement, "If I had had to write
in English from the beginning instead of writing in Japanese first and then
using MT, I probably would have chosen a simpler topic," 40% (6 students)
answered, "Yes," and 33% (5 students) answered "Somewhat yes.
The quality of the Japanese essays
in that class was remarkably high, in general. The post-editing goal for the
last five weeks was to make the English translation match that high quality.
Unfortunately, due to my lack of instructional competence, the final revision
was not satisfactory. Improvement of teaching skills is necessary, considering
that in the questionnaire, 73% (11 students) answered "yes" and 20%
(3 students) answered "somewhat yes" to the prompt "I would like
to use MT for various occasions in the future actively.
Although there are still issues to
be addressed, I believe that the introduction of machine translation will
increase the quality of writing instruction, such as academic essay writing of
1,000 English words or more.
Conclusion
I can summarize the above
discussion in the following six points.
(1) AI allows learners to continue
to think while writing drafts in their native language at their intellectual
level, rather than in the active vocabulary of a foreign language, which vastly
underrepresents their intelligence.
(2) AI increases opportunities in
which learners can write about topics at a level appropriate to their
intellectual interests. Those changes probably have a positive impact on
learners' self-esteem.
(3) The AI expresses what learners
intended, albeit imperfectly, at their foreign language comprehension
vocabulary level. In other words, those expressions offer learners an
opportunity to think and write on a level far higher than their current level
of active vocabulary.
(4) If the English output of the AI
is beyond learners' comprehension, learners cannot benefit from the AI's
English. However, if their vocabulary level is adequate for understanding the
output, learners can discover that many specific expressions that the AI presents
can become part of their productive capacity. AI’s English becomes a unique learning
material for reading and writing.
(5) If the students learn to read
the English output of the AI critically and can revise AI’s English from their
critical assessment, they will gain an in-depth understanding of their comprehension
vocabulary. In addition, they will learn ways to apply their comprehension
vocabulary to express their ideas. Their English writing skills will increase
simultaneously with their reading abilities.
(6) In summary, writing-assistant
AI enables learners to think and express themselves using their passive
vocabulary instead of their active vocabulary in the target language (English).
It also promotes the transformation of words for understanding into words for
expression.
For these reasons, I believe that
instructors can wisely introduce AI in writing classes where learners need to
express their complex thoughts in a long essay precisely. AI can benefit
learners both in terms of their use of English and their English learning.
However, learners need critical reading literacy to understand English that
expresses topics at their intellectual level.
Lastly,
I wrote this essay rather
impulsively, inspired by a conversation I had yesterday with my colleagues.
Since my mind was not sufficiently organized for the topic, I had to draft fast,
scan I, and rewrite to think coherently. Therefore, my language choice was my
native language Japanese, my best thinking tool. It took about an hour and a
half to complete the Japanese version.
I input my Japanese text into
DeepL, copied its English output into Grammarly, and revised the English on the
Grammarly screen. When the expression seemed dull, I explored the possibility
of rewriting it in Wordtune. I used DeepL again for translating my revised
English into Japanese to check if it might contain ambiguous expressions. I rewrote
this way because I wanted to share my thoughts with my native English-speaking
colleagues in the best way I can. It took about two hours and a half to produce
the English translation.
Without AI, I would have given up
on English translation because it would have taken too much time. Also, if I
had written this essay in English initially, my thoughts would have been more
confused than they are now, and my expression would have been much weaker.
However, AI empowered me to express my thoughts fast and communicate them to my
colleagues in English. During the revision process, I also learned more about
English expression.
From experiences in my classrooms
and workplace, I believe that AI can effectively promote students' use and
learning of English. AI can empower students and make them more autonomous.
Of course, one can use any tool
foolishly.
Posts
