Evidence of cultural differences between American and Japanese mainstream science and engineering contexts from analysis of classroom discourse

ABSTRACT

Teaching styles in science and engineering instruction were compared by analysing corpora of transcripts of lectures delivered in English and Japanese at leading universities in the United States and Japan, respectively. Our findings were compatible with cultural differences related to power distance and field dependence, which have been reported in the literature. Teaching styles seem to simultaneously result from the cultural context as well as reinforce it. Science and engineering instruction in the Japanese educational context tends to reflect and reinforce a personalised transmission of knowledge style, while instruction in the American context tends to match and reinforce learning styles characterised by impersonal, inductive thinking.

KEYWORDS:

• Corpus of lectures
• education and culture
• language and pedagogy

Introduction

Adoption of English as the medium of instruction (EMI) in higher education is expanding rapidly throughout the world. The Nordic and Baltic countries have a large number of EMI educational programmes per 100,000 inhabitants (Hultgren et al. 2015), and in spite of the potential dangers, such as impoverishment of local languages, associated with adoption of EMI (especially when the local population is small: Brock-Utne 2001), other cultures where English is not the mainstream language are also adopting EMI in higher education: from Europe (Doiz, Lasagabaster, and Sierra 2013; Wachter and Maiworm 2014; Earls 2016) to Asian countries and regions such as Malaysia (Tan, Saw, and Lan 2011), Hong Kong (Evans and Morrison 2011) and, more recently, Japan (Bradford and Brown 2018). EMI educational programmes are expected to attract international students, thus forming classrooms having multicultural and multilingual backgrounds.

In order to alleviate the linguistic challenges faced by instructors and/or learners who are non-native speakers of English (NNS) and become involved in EMI, we have created OnCAL (Online Corpus of Academic Lectures, http://www.oncal.sci.waseda.ac.jp/). We hypothesised that a corpus (a database of texts) containing many samples of lectures delivered in English would allow analysis of the linguistic features of such lectures and thus enable instructors and students in science and engineering to become more familiar with this genre of English usage (Kunioshi et al. 2016). OnCAL can raise the awareness of instructors and students of the relationship between the aspects of classroom language and pedagogical efficacy. The main purpose of building OnCAL was therefore to present concrete examples of utterances delivered by instructors at two leading American universities (mostly native speakers of English, NS), as a reference to prepare for teaching in English or listening to lectures delivered in English. We were able to identify expressions frequently used to realise various pedagogical functions, e.g. explaining cause and effect or proposing thought experiments, to allow OnCAL users to easily find examples of utterances for such pedagogical purposes.

However, language is not the only important factor to consider in formal education situations; the learning and teaching styles as well as the background culture should also be taken into consideration. In the United States, the instruction offered in mainstream education (science and engineering education included) has been criticised for paying little attention to the knowledge and orientations that students have acquired before entering formal education. According to Medin and Bang (2014), ‘children come to school with knowledge, orientations, values and practices that are relevant to science learning and that reflect their own culture’ and, citing Bell et al. (2009), ‘when these orientations are supported, students are more engaged, identify with and are more successful with science than when these orientations are ignored or discouraged’ (5). In addition, Felder and Brent (2005) suggest that many good students drop out from engineering programmes because of dissatisfaction with the instruction they receive as it does not match their expectations and orientations; these authors maintain that ‘if completely individualized instruction is impractical and one-size-fits-all is ineffective for most students, a more balanced approach that attempts to accommodate the diverse needs of the students in a class … is the best an instructor can do’ (57).

For the present work, we hypothesised that there would be differences between instruction practices in science and engineering education offered in American and Japanese mainstream university contexts. The purpose of our present work was to identify specific features that would reflect the differences between the cultural contexts. To test our hypothesis, we examined corpora of transcripts of lectures in science and engineering delivered in English at two leading American universities and in Japanese at two major Japanese universities.

Learning styles, teaching styles and culture

One of the major theories of learning that is propelling recent work in education is based on the concept of situated learning (Lave and Wenger 1991; Flowerdew 2013). This concept is rooted in sociocultural theory or the influence of culture on cognitive development. The learning of a child, according to Vygotsky (1980, 2012), occurs via internalisation of experiences of interactions with adults, who take the child to the zone of proximal development to make further development possible. The ways adults help children are transmitted through the generations by observation or by verbal expression, and become procedural, some specific to a culture. Learning is also considered to result not only from purposeful interventions of adults or masters but also from participation in communities of practice (Wenger 1999). More importance is now given to knowledge acquired in everyday life, through participation in different communities ranging from the home or neighbourhood to the workplace (Lave and Wenger 1991). Everyday settings are therefore sources of knowledge that learners develop, independently of schooling. The knowledge provided by experience in everyday or culturally specific settings tends though to be procedural rather than conceptual (Inagaki 1982). This may explain why the knowledge obtained in informal settings can facilitate or sometimes inhibit the learning process in formal educational settings. In other words, the ability to transfer knowledge obtained in a context to another, for example from everyday life to school and vice versa, is fundamental for effective learning to take place (Bransford, Brown, and Cocking 2000).

In an effort to integrate theory and practice, Lewin (1946) developed techniques of action research and laboratory training, and suggested a learning cycle of planning, execution (action) and fact-finding for the evaluation of the action, and the evaluation serves as a basis for planning the next step. Dewey (1963) also emphasised the importance of connecting education to everyday life; in a cycle similar to Lewin’s, Dewey suggested that learning transforms impulses and desires elicited by concrete experience into purposeful action in the real world. For Piaget, learning also comes from the interaction between the individual and the environment, and development requires moving from a phenomenal and active-egocentric to an abstract-constructionist and reflective-internalized mode of knowing, in an interaction between processes of accommodation of concepts to concrete experiences and assimilation of new concrete experiences into existing concepts (Wadsworth 1996). These sociocultural views of cognitive development acknowledge the importance of actual experiences in learning. The experiences of a learner are necessarily situated, and the cognitive development of learners in different places or cultures will be influenced differently, potentially leading to the formation of different learning styles.

Learning styles, which are related to the knowledge, orientations or preferences that learners acquire through experiences in both formal and informal contexts, have been discussed extensively. Coffield et al. (2004a, 2004b) have identified 71 models of learning styles, carefully analysed the most influential 13 and concluded that none of the models can be used as a theoretical justification for changing practice, at least in higher education. However, the Experiential Learning Theory (ELT) of Kolb (2015), one of the 13 most influential models in the reviews of Coffield et al. (2004a, 2004b), has attracted considerable attention in engineering education, as a basis for the promotion of active learning (Christie and de Graaff 2017). ELT adopts a learning cycle (Figure 1) that incorporates Lewin’s cycle, the accommodation and assimilation processes of Piaget, and the transformation of impulses into the purposeful action described by Dewey. The relation of ELT to active learning can be seen in the integration of observation, reflection, abstraction and action, in the dual dialetics of action–reflection and of concrete–abstract. ELT predicts that a learner deepens understanding through the following cycle: concrete experience (CE) → reflective observation (RO) → abstract conceptualisation (AC) → active experimentation (AE) → new experience. Zull (2002) finds evidence for the validity of the ELT learning cycle by relating the four stages to four different functions of the brain, which brain imaging methods have shown to be localised in different zones: when sensing a concrete experience (e.g. when seeing a word) the sensory cortex of the brain is activated, the integrative cortex at the back of brain is activated upon reflective observation (e.g. when hearing a word), the frontal integrative cortex is activated by abstract thinking (e.g. when generating verbs) and the motor brain is related to active experimentation (when speaking a word).

Figure 1. Learning cycle of the Kolb model, adapted from Kolb (2015, 68).

According to ELT, a learning style results from preference of one stage in the cycle over the others: converging (AC → AE, preference for practical uses of ideas and theories), diverging (CE → RO, preference for observation rather than action), assimilating (RO → AC, preference for inductive reasoning) and accommodating (AE → CE, preference for action or hands-on experimentation) styles. The Learning Style Inventory (LSI) is a measurement device designed by Kolb (1985) to determine the main style of a learner. Cagiltay (2008) used LSI in Turkey with 285 engineering students and found that the percentages of students having converging, diverging, assimilating and accommodating styles were, respectively: 24%, 24%, 47% and 5%. These results differ from those obtained from the use of LSI in the United States with 1013 engineering students (Sharp 2001): 40%, 8%, 39% and 13%. Both reports are in agreement about the assimilating style representing a large portion of engineering students, but the difference in the share of the divergent style is significant. These findings indicate that the way engineering students learn may depend significantly on their cultural background, and that learners strong in one stage of the learning cycle may lack skills related to the other stages.

Another approach to cognitive learning styles is that of Witkin et al. (1977), based on field dependence (FD) or field independence (FID) styles. The basic idea is that FD learners are strongly influenced by the immediate environment or social context, while FID learners can abstract items and themselves from the context, and have more analytic skills than FD learners. Field-dependent learners need to develop field independence to learn effectively (Pithers 2002).

Turning to concepts of culture, one major influence has been that of Hofstede, Hofstede, and Minkov (2010, 6) who define ‘culture’ as ‘the collective programming of the mind that distinguishes the members of a group or category of people from others’. Lewis (2006) explains that this collective programming is not inherited but ‘begins in the cradle and is reinforced in school and in the workplace’ (17). He states that respect for elders and traditions is an important item included in the Japanese collective programming (18). Formal instruction in schools (including universities) is therefore one of the factors that can form, modify or at least influence the learning styles of students. In turn, formal instructional practices, or teaching styles, can be expected to reflect the cultural context in which the instruction is delivered.

Knowing the style of a learner can offer insights on how to provide instruction that can respond to learner needs. For example, through proper pedagogical discourse (examples and analogies) that relates the concepts being taught to everyday experience, formal instruction can refine the meanings that learners have assigned to an object of learning from everyday experience alone (Tsui 2004). In addition, while culture influences individuals, it can be changed by individual actions (Rogoff 2003). Therefore, knowing about educational practices in situated contexts should be the first step towards improving the efficacy of education.

Corpus building

The transcripts of lectures delivered through English were downloaded from MIT Open Course Ware (MIT OCW, https://ocw.mit.edu/ and from Stanford Engineering Everywhere (SEE, https://see.stanford.edu/) The transcripts were uploaded to a database, and expressions frequently used to realise specific pedagogical purposes were identified. The creation and implementation of OnCAL (Online Corpus of Academic Lectures, http://www.oncal.sci.waseda.ac.jp/) has been described in Kunioshi et al. (2016). OnCAL comprises transcripts of 430 lectures that correspond to 395 hours of class time, as of February 2017.

Video recordings of 44 lectures delivered through Japanese by a professor of information science in a major Japanese private university were transcribed and uploaded to a new version of OnCAL (hereafter OnCAL.jp) that allows analysis of Japanese texts. The 44 lectures correspond to 32 hours of class time and were from 3 courses related to information science that were delivered to undergraduate students. Another set of lectures that we transcribed was made available online by a major national Japanese university. Most of the courses to which these lectures belonged were delivered by a team of professors to undergraduate students. The 21 recordings correspond to 33 hours of class time, and the transcripts were uploaded to OnCAL.jp. The transcription work is still in progress for other courses, and six of these, corresponding to 7 hours of class time, have already been included in the corpus. The total number of lectures in OnCAL.jp is thus 71, which corresponds to a total lecture time of 72 hours.

OnCAL and OnCAL.jp allow evaluation of the frequency with which a word or expression is used in the lectures and can also show how the word/expression is used in specific cases. Figure 2 shows the concordance lines for ‘instructor’ in OnCAL. The results in the figure are sorted according to (1) word preceding the word containing the search string, (2) word containing the search string and (3) word following the word containing the search string, so that all instances of the pattern ‘your recitation instructor’ appear together and can be counted easily. The sort order can be changed according to the type of pattern to be found. If a search string is selected appropriately and the results sorted suitably, the frequency and patterns of usage of terms can reveal linguistic features that are related to the cultural context of the lectures.

Figure 2. OnCAL display of results for the search term ‘instructor’.

Results and discussion

Careful choice of terms to be searched can reveal differences between American and Japanese mainstream cultures and also among science and engineering education contexts.

We first chose pronouns as search terms, hypothesising that searches for ‘I, you, we’, etc. and the correspondent Japanese terms would reflect features of the respective cultural backgrounds. The frequencies of first and second person pronouns, and their genitive and dative cases are listed in Table 1. First, it should be noted that for the Japanese pronouns, only the most frequent ones are shown in the table, but all the variations were counted: ‘boku’ and ‘ore’ are included in the count for ‘watashi’, ‘wareware’ includes ‘watashitachi’ and ‘warera’, while ‘kimitachi’, ‘kimira’, ‘anatagata’ and ‘shokun’ are included in ‘minasan’. In addition, the Japanese genitive case is formed by addition of the particle ‘no’, and the dative case by addition of the particle ‘ni’. In English, ‘you’ includes the dative case, in singular and plural. We can see that the total number of pronouns is much higher in lectures delivered in English than in Japanese. This is because in English, it is common to find sentences such as ‘I’m going to explain to you now what we physicists mean by … ’ (from a physics lecture) which contains three pronouns. In Japanese, pronouns are frequently not needed neither as subject nor complement. For example, in ‘koreha nagasa L, menseki A to shimashou’ (from a lecture on semiconductors) which means ‘let us call this length L and area A’, ‘shimashou’ stands for ‘let us’ with no need for a pronoun or even an explicit subject. The other significant difference is that the second person (singular and plural) pronouns are the most frequent, and the pronouns related to first person plural is the least frequent in the lectures delivered in English. This result can be seen as evidence of the interactive character of the lectures delivered in the United States: instructors are always directing their utterances towards the audience. In contrast, in Japanese, pronouns for the first person singular are the most frequent with the first person plural and second person plural showing almost the same frequency. This serves as evidence of teacher-centred instruction.

Table 1. Frequencies of usage of first and second person pronouns, with their genitive and dative cases, in both corpora of lectures delivered in English (OnCAL) and Japanese (OnCAL.jp).

Pronoun	Frequency in OnCAL (430 lectures)	Total for first and second person	Average per lecture in OnCAL	Word	Frequency in OnCAL.jp (71 lectures)	Total for first and second person	Average per lecture in OnCAL.jp
I	65,000*	75,523	151.1	watashi	775	1,002	10.9
my	5,000*		11.6	watashino	188		2.6
me	5,459		12.7	watashini	39		0.5
mine	64		0.15
we	39,580	44,686	92.0	wareware	343	494	4.8
our	3,386		7.9	wawareno	128		1.8
us	1,695		3.9	warewareni	23		0.3
ours	25		0.06
you	89,876	97,554	209.0	minasan	399	464	5.6
your	7,632		17.7	minasanno	57		0.8
				minasannni	8		0.1
yours	46		0.1
Total	217,763		506.4	Total	1,960		27.6

*The actual number of hits for ‘I’ and ‘my’ were 66,462 and 5,187, respectively. However, ‘I’ are ‘i’ are used at least 1,000 times, and ‘m_y’ at least 100 times as symbols, and this inflates the frequency for pronouns. The values 65,000 and 5,000 are therefore estimated frequencies.

Another Japanese word that we examined was ‘sensei’, which can be translated into English as ‘teacher’. A university ‘professor’ would be referred to as ‘kyoju’, and an ‘instructor’ as ‘kyoin’ in Japanese. However, ‘sensei’ is mainly used together with a name, and in this sense, its usage is similar to the usage of ‘professor’ in the lectures delivered through English. Table 2 shows the frequency with which ‘sensei’, ‘kyoju’ and ‘kyoin’ are used in OnCAL.jp, and those for ‘teacher’, ‘professor’ and ‘instructor’ in OnCAL. It should be noted that OnCAL.jp comprises only 71 lecture transcripts and is therefore much smaller than OnCAL, which comprises 430 transcripts. Because some of the courses delivered in Japanese were taught by a team of professors and this inevitably forces the instructor of a session to frequently cite the instructor(s) of previous session(s) of the course, the frequency of the words in OnCAL.jp becomes inflated. However, detailed observation of each hit showed that the citations of other instructors of the same course represent at most one-third of the 313 hits. Even considering that half of the 313 hits were just citing other instructors of the same course, the average frequency per lecture of ‘sensei’ is about 2.2, which is much larger than the total average of about 0.5 times per lecture for ‘teacher’, ‘professor’ and ‘instructor’.

Table 2. Frequencies of usage of the words ‘sensei’, ‘kyoju’ and ‘kyoin’ in OnCAL.jp, and ‘teacher’, ‘professor’ and ‘instructor’ in OnCAL.

Word	Frequency in OnCAL.jp (71 lectures)	Average per lecture in OnCAL.jp	Word	Frequency in OnCAL (430 lectures)	Average per lecture in OnCAL
sensei	313	4.4	teacher(s)	25	0.06
kyoju	23	0.3	professor(s)	159	0.37
kyoin	9	0.1	instructor(s)	35	0.08
Total	345	4.8	Total	219	0.51

According to Hofstede, Hofstede, and Minkov (2010, 59), the Power Distance Index (PDI) of Japan is 54, which is higher than the 40 of the United States. In large power distance cultures, ‘the educational process is teacher centred’ and ‘highly personalised’: especially in more advanced subjects at universities, what is transferred is seen not as an impersonal “truth,” but as the personal wisdom of the teacher' (69). The high frequency of ‘sensei’ can therefore be attributed to the fact that in the Japanese educational context, instruction not only covers scientific truth but also tends to include the people who discovered, developed or support that truth. This personalised transmission of knowledge is related to the respect for elders and to traditional transmission of skills from master to apprentice. The example ‘ … wo yogen shita no ha Yukawa Hideki sensei desu’, which means ‘it was Professor Hideki Yukawa who proposed that … ’, is a typical utterance where the instructor mentions an illustrious scientist in a way that gives more emphasis on the scientist than on his/her discovery. In English, cases similar to ‘ … were later discovered by Hertz, and that was a great victory for the theory’, which puts more emphasis on the scientific fact/theory, are more frequent. It should be noticed that the mention of illustrious scientists such as Hertz or Einstein does not come together with ‘professor’ in English or ‘sensei’ in Japanese. For example, a search for ‘Maxwell’ in OnCAL generates 116 hits, but 61 of these were not personal references to Maxwell, like in ‘Here is the Maxwell–Boltzmann distribution … ’, and the valid personal references to Maxwell, like in ‘But Maxwell believed that vacuum in a way behaves like … ’ amounts to 55, coming from 19 sessions (three different instructors). In OnCAL.jp, a search for ‘Einstein’ leads to 155 hits, of which 109 are personal references made in 16 sessions (9 different instructors), and 46 involve technical terms such as ‘Bose-Einstein gyoshukutai’ or ‘Bose-Einstein condensates’. The personal references per session are therefore more frequent in the lectures delivered in Japanese. Japanese culture therefore seems to influence the instruction style, which, in turn, can be seen as reinforcing a field-dependent learning style in students through a teacher-centred approach.

In a culture where the power distance is small, the educational process is ‘student centred’ and ‘is rather impersonal: what is transferred are “truths” or “facts” that exist independently of this particular teacher’ (Hofstede, Hofstede, and Minkov 2010, 69–70). This does not contradict the expectations elicited by the results of Sharp (2001) which showed that engineering students in the United States have mainly the converging (40%) and assimilating (39%) learning styles of the Kolb (2015) model, because students having these styles are comfortable with the practical use of ideas and theories and with inductive reasoning, respectively. Instruction in the American context may therefore reflect the impersonal or individualised critical logical thinking style typical of the Western cultural context and at the same time reinforce some learning styles while neglecting others derived from views of nature different from the typical mainstream American or Western view (Medin and Bang 2014, 79–83). Different groups of people, e.g. Hispanics, African Americans, Native Americans, Hawaiians, European Americans and Middle Eastern Americans, have been reported to display different learning styles (Bowers and Flinders 1990, 73). Learning styles similar to those of Native Americans were also observed in Aboriginal Australian university students (Boulton-Lewis, Marton, and Wills 2001). Students having different learning styles may have different needs, and effective instruction can be realised by taking these differences into consideration.

Further evidence of cultural differences between the Japanese and American educational contexts comes from the number of student interventions. The number of clear student interventions found in OnCAL and OnCAL.jp were 2,737 and 19, respectively. On average, there are 6.4 student interventions per session in the lectures delivered in English. Some of the lectures delivered in Japanese were from online courses, and consequently do not allow for student intervention, but this cannot justify the extremely small number of student interventions in the lectures delivered through Japanese. Furthermore, the few student interventions were actually responses elicited by questions from the instructor. In other cases, students wrote questions in reports submitted during the course, and the instructor answered those questions in the classroom, with no need for students to ask a question aloud during a session. In another case, at the end of a session, the instructor invited ‘those who may have questions’ to remain in the classroom after the session finished. This again confirms the prediction of Hofstede, Hofstede, and Minkov (2010) that in a culture where the power distance is large, ‘teachers outline the intellectual paths to be followed’, ‘students in class speak up only when invited to’, ‘with the teacher initiating all communication’ (69). This aspect of the Japanese classroom is again typical of a personalised transmission of knowledge, being more field dependent than in the American context. Obviously, the ‘silence’ of Japanese students does not necessarily mean that they cannot engage in active learning. Instructors can elicit active learning by posing questions to students, inviting them to observe, make abstractions, perform thought experiments (Stephens and Clement 2012) and guess answers before providing the solutions, without forcing the students to talk.

Our findings show that instructor discourse does reflect the cultural context in which the instruction is delivered. Analysis of the corpora of lecture transcripts revealed concrete differences of the cultural contexts, as well as details concerning the teaching styles in the American and Japanese contexts. Awareness of these differences is important for implementing methods for improving instruction and learning. As the Japanese corpus is much smaller than the English corpus, efforts are being made to expand it. In addition, more words/expressions should be investigated for a more comprehensive comparison.

Conclusion

Corpora of transcripts of science and engineering lectures delivered in English and Japanese were analysed to find concrete examples of utterances that reveal differences in cultural contexts. The science and engineering instruction in the Japanese universities reflected the teacher-centred, personalised transmission of knowledge that characterises the field-dependent style of the Japanese cultural context, and may be reinforcing this style in the students. On the other hand, the instruction in the American universities was seen to match a learning style related to inductive reasoning which is typical of the mainstream American culture.

The Japanese part of the corpus needs to be expanded and more words or expressions should be analysed for a more comprehensive comparison to identify other concrete differences and suggest ways to improve instruction in the classroom.

Acknowledgement

This research was supported by a Grant-in-Aid for Scientific Research (B), grant number 16H03068, from the Japan Society for the Promotion of Science (JSPS).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Nilson Kunioshi is an engineer and professor in the Graduate School of Creative Science and Engineering of Waseda University. He teaches and conducts research on chemical kinetics and dynamics, and also teaches technical writing for research articles to science and engineering students. He participated in the implementation of an English-medium science and engineering programme in Waseda University.

Judy Noguchi is a professor emerita of Kobe Gakuin University, Faculty of Global Communication. She currently teaches presentation and writing skills for undergraduate and graduate students majoring in science, engineering and medicine and related fields at various universities. She is interested in the use of language in knowledge construction and effective language learning and teaching methods for ESP (English for specific purposes).

Kazuko Tojo is a native Japanese speaker and professor at Osaka Jogakuin University, where she teaches English for Academic Purposes (EAP) and ESP. Her research interests include genre-based curriculum and textbook development for Japanese ESL students. She is also involved in English teacher training programmes.

ORCID

Nílson Kunioshi http://orcid.org/0000-0001-7608-179X

Additional information

Funding

This work was supported by Japan Society for the Promotion of Science [Grant Number 16H03068].