Language Awareness

ISSN: 0965-8416 (Print) 1747-7565 (Online) Journal homepage: https://www.tandfonline.com/loi/rmla20

The effects of input and output based instruction

on the development of L2 explicit and
automatised explicit knowledge: a classroom
based study

Tamanna Mostafa & YouJin Kim

To cite this article: Tamanna Mostafa & YouJin Kim (2020): The effects of input and output based
instruction on the development of L2 explicit and automatised explicit knowledge: a classroom
based study, Language Awareness, DOI: 10.1080/09658416.2020.1760292

To link to this article: https://doi.org/10.1080/09658416.2020.1760292

Published online: 20 May 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=rmla20
Language Awareness
https://doi.org/10.1080/09658416.2020.1760292

The effects of input and output based instruction on the

development of L2 explicit and automatised explicit
knowledge: a classroom based study
Tamanna Mostafa and YouJin Kim
Department of Applied Linguistics and English as Second Language, Georgia State University, Atlanta, GA, USA

ABSTRACT ARTICLE HISTORY

The study investigates to what extent English as a second language Received 26
(ESL) learners develop explicit and automatised explicit knowledge of December 2018
two target-forms (third person singular ‘s’ and future conditional Accepted 2 April 2020
if-clause) after input and output based instruction and whether learning KEYWORDS
effects differ depending on the target forms. The study follows a pre- Explicit knowledge;
test–post-test design. Seventy-two intermediate level ESL learners from automatised explicit
three intact classes were randomly assigned to two experimental knowledge; input based
groups (input and output) and one control group. The experimental instruction; output based
groups completed a pre-test and received relevant instructional treat- instruction; classroom
ment: the input group received processing instruction that did not based research; second
language acquisition
involve any production of output, whereas the output group received
output based instruction. Immediately after the instructional treatment,
the experimental groups took a post-test and after two weeks, a delayed
post-test. The control group took the same tests without receiving any
instructional treatment. An error correction test and an oral narrative
production test were used to measure explicit and automatised explicit
knowledge, respectively. The results showed that both the input and
output groups outperformed the control group in developing explicit
knowledge of the target forms, and the output group outperformed
the input group in developing automatised explicit knowledge of both
target forms.

Introduction
The relationship between different components of second language (L2) knowledge and
various instructional types has been one of the main interests in Instructed Second Language
Acquisition (ISLA) (Loewen & Sato, 2017). Of central importance to second language acqui-
sition (SLA) is how learners acquire L2 explicit (that involves conscious awareness of linguistic
norms) and implicit (involving unconscious or intuitive knowledge of those norms) knowl-
edge (Ellis, 2005; Hulstijn, 2005; Suzuki & DeKeyser, 2015). To discuss knowledge types, the
present study adopts skill acquisition theory (DeKeyser, 2015), which draws a distinction
between declarative and procedural knowledge. The dichotomy of explicit/implicit

CONTACT Tamanna Mostafa [email protected]

Supplemental data for this article is available online at https://doi.org/10.1080/09658416.2020.1760292.
© 2020 Informa UK Limited, trading as Taylor & Francis Group
2 T. MOSTAFA AND Y. KIM

knowledge often overlaps to a great extent with that between declarative/procedural knowl-
edge, although the distinctions between these two pairs of terms may not always coincide
(DeKeyser, 2009, 2015).
In the present study, declarative knowledge is equivalent to explicit knowledge, which
refers to conscious knowledge about language rules (DeKeyser, 2015). When explicit knowl-
edge of rules turns into a behaviour, it becomes procedural knowledge, that is, ‘knowledge
about how to perform various processes or behaviors’ (DeKeyser & Criado, 2012, p. 1; see
also DeKeyser, 2009, 2015). Although this proceduralization of knowledge is not time-con-
suming, learners need a large amount of practice to use language with ‘complete fluency or
spontaneity, rarely showing any errors’ (DeKeyser, 2015, p. 95), and it leads to automaticity
in language use ‘where knowledge is completely procedural’ (ibid, p. 96). According to Suzuki
and DeKeyser (2015), ‘fully automatized knowledge may be used with little or no awareness
and would therefore be nearly impossible to tease apart from implicit knowledge through
behavioural measures’ (p. 864). However, automaticity is a gradual process, and ‘even highly
automatized behaviors are not 100% automatic’ (DeKeyser, 2015, p. 96). Therefore, in recent
empirical studies, automatised explicit knowledge, defined ‘as a body of conscious knowl-
edge involving different levels of automatization’ (Suzuki & Dekeyser, 2017a, p. 751), has
been argued to be a distinct construct from implicit knowledge, which is ‘knowledge without
awareness’ (Suzuki & Dekeyser, 2017a, p. 751). Thus, although both automatised explicit
knowledge and implicit knowledge are accessed rapidly under time pressure, the former
involves awareness of the target forms, whereas the latter does not (Suzuki, 2017).
Although there have been laboratory based experimental studies on developing explicit
or implicit knowledge of artificial languages (e.g. Andringa & Curcic, 2015; Cintron-Valentin
& Ellis, 2015), those studies might not be very informative in ISLA context, as the findings
do not provide direct pedagogical implications. In contrast, studies focussing on English as
second language (ESL) learners’ acquisition of target L2 forms in a classroom based context
may add to the ecological validity of the investigations. While there have been empirical
studies examining the effects of different instructional variables on L2 receptive and pro-
ductive knowledge (e.g. Erlam, 2003; Morgan-Short & Bowden, 2006; Toth, 2006), very few
studies (Erlam et al., 2009; Loewen & Erlam, 2009; Reinders & Ellis, 2009) investigated learners’
development of implicit and explicit knowledge from instructional treatments. However,
the time-pressured tests used in those studies to measure implicit knowledge, such as elicited
imitation test (EIT), have been suggested in recent research as measures of automatised
explicit knowledge (Suzuki, 2017; Suzuki & DeKeyser, 2015; Vafaee et al., 2017). Thus, there
is hardly any clear finding on the kind of knowledge learners develop from classroom based
instructions. Given the difficulty of measuring implicit knowledge even in controlled labo-
ratory contexts, it may be more relevant in an instructional context to examine whether
learners can attain some degree of automaticity in using the L2 forms of which they have
consciously gained explicit knowledge through formal instruction (DeKeyser, 2015).
Therefore, the present study examines the effects of two types of instructional treatments:
input based (operationalised as Processing Instruction) and output based (Erlam et al., 2009)
on learners’ development of explicit and automatised explicit knowledge of a morphological
(third person singular ‘s’) and a syntactic (future conditional if-clause) target form (Suzuki &
DeKeyser, 2017a).
 Language Awareness 3

Literature review
Key concepts
Different types of L2 knowledge
Whereas Krashen (1999) believed that explicit knowledge of rules can never lead to implicit
or acquired knowledge (non-interface position), Ellis (2007) argued that explicit knowledge
of grammar rules as well as memory of patterns facilitates ‘conscious creation of utterances
whose subsequent usage promotes implicit learning and proceduralization’ (p. 32).
Furthermore, from a strong interface position, DeKeyser (2009) said that with enough prac-
tice, explicit knowledge leads to ‘proceduralization, (at least partial) automatization, and
only in some cases, eventually, after proceduralization and automatization have taken place,
to implicit representation’ (DeKeyser, 2009, p. 127). However, complex skills such as L2 learn-
ing may not always become fully automatic because automatisation is a long and gradual
process (DeKeyser, 2015; Suzuki, 2017). Therefore, in recent research, automatised explicit
knowledge, wherein learners might still draw on awareness of L2 forms in rapid execution
of language skills under time pressure, is distinguished from implicit knowledge, wherein
linguistic skills are used without awareness (Suzuki, 2017). Automatised explicit knowledge
was also found to be predictive of implicit knowledge in a naturalistic L2 acquisition context
(Suzuki & DeKeyser, 2017a). In contrast, conscious awareness of L2 rules usually stems from
explicit instruction of those rules (DeKeyser, 2015; Jiang, 2007).

Input and output based language instruction

Previous empirical studies compared the effects of comprehension based instruction (CBI)
and production based instruction (PBI) on L2 learning (Shintani et al., 2013). One approach
to CBI is Processing Instruction (PI) based on VanPatten’s (2015) model of input processing.
The input that learners pay attention to and where form-meaning connections are also made
turns into intake, which affects the restructuring process, and thus, new knowledge is incor-
porated into the developing system (Wong, 2004). As input based instruction, the present
study focuses on PI that consists of explicit explanation of target form rules, instruction to
avoid the wrong processing strategy, and structured input activities (Wong, 2004).
On the contrary, PBI elicits correct production of target forms from learners (Shintani et al.,
2013). In the present study, the term PBI is used synonymously with output based instruction
(OBI) that involves explicit explanation of target form rules and output based activities (Erlam
et al., 2009). OBI is motivated by the belief that producing output pushes learners to ‘notice’
the gap in their current interlanguage and may also engage them in testing their hypotheses
about the L2 forms that can ultimately lead to L2 knowledge (Swain, 2000, p. 102).

Empirical studies on the effectiveness of PI and OBI

A number of studies have compared the effectiveness of PI and OBI on learners’ development
of L2 receptive and productive knowledge. Studies such as Toth (2006) and Morgan-Short
and Bowden (2006) found better performance of the output group participants in production
based tasks among adult learners of Spanish as L2 (average age 19.3 years in Morgan-Short
& Bowden, 2006) in US university contexts. Similarly, Erlam (2003) found greater gains for
the OBI group compared to PI group in developing receptive and productive knowledge of
4 T. MOSTAFA AND Y. KIM

direct object pronouns in L2 French among secondary level school students (average age
14 years) in New Zealand. On the contrary, Lee and Benati (2007), which was conducted with
undergraduate level L2 learners of Italian and Greek in a British university and Benati (2005),
conducted with secondary level learners of English in China and Greece, found that L2 learn-
ers who received PI performed better than those receiving OBI in developing receptive and
productive knowledge of the target L2 forms. Additionally, in Farley (2004) and Collentine
(1998), both PI and OBI groups were equally effective in developing receptive and productive
knowledge of Spanish subjunctive in US university contexts with no significant difference
between the performances of the two groups. These studies mostly used multiple choice
listening comprehension and written sentence completion tasks for measuring L2 receptive
and productive knowledge, respectively (except Erlam, 2003, who used an oral picture
description task for measuring L2 productive knowledge). Therefore, these previous studies
mostly used tasks (e.g. multiple choice, sentence completion) that tap into learners’ conscious
knowledge of L2 forms (i.e. explicit knowledge) to investigate the effects of different instruc-
tional treatments on L2 learning. In a meta-analysis of similar studies, Shintani et al. (2013)
found that CBI resulted in greater receptive knowledge than PBI only in the short term,
whereas PBI had greater sustained effect on productive knowledge in the long term. Shintani
et al. (2013) further suggested that CBI helps develop only explicit knowledge, which is less
durable than implicit knowledge (R. Ellis, 2004), whereas PBI facilitates developing implicit
knowledge. However, such interpretations of the findings, as mentioned by Shintani et al.
(2013), are ‘speculative’ (p. 321) because very few studies examined the effects of L2 instruc-
tions on the development of different types of L2 knowledge (e.g. explicit and implicit).
Loewen and Erlam (2009) found that there was no development of L2 explicit and implicit
knowledge of third person singular ‘s’ from incidental and extensive exposure to it during
input based instruction in adult ESL classrooms in New Zealand. On the contrary, Erlam et al.
(2009), which was also conducted in a similar ESL context, focussed on the relative effective-
ness of two instructional methods, OBI and PI, on developing the implicit and explicit knowl-
edge of English indefinite articles. After the instructional treatment, the OBI group had
greater gains than the control group in developing implicit and explicit knowledge of the
target form, but Erlam et al. (2009) did not find any significant difference between the OBI
and the PI groups in this regard. Both Erlam et al. (2009) and Loewen and Erlam (2009) used
an untimed grammaticality judgement test (UGJT) to measure explicit knowledge and a
version of EIT to measure implicit knowledge. In the EIT, test-takers first listened to a set of
statements, indicated whether they agreed or disagreed with each, and then, orally repeated
those statements. As the test-takers performed the EIT under time pressure, they were sup-
posed to rely on feel and focus on meaning (Ellis, 2009). Additionally, in the UGJT, the par-
ticipants had to judge the grammaticality of a set of sentences under no time pressure (Ellis,
2009). However, it was found in recent research that EIT, used to measure implicit knowledge
in Erlam et al. (2009) and Loewen and Erlam (2009), taps into automatised explicit knowledge
rather than implicit knowledge (Suzuki & DeKeyser, 2015). Furthermore, unlike Erlam et al.
(2009), Lindseth (2016) investigated the effects of only form-focussed instruction (FFI) involv-
ing output based activities on the acquisition of subject-verb inversion in German. After the
instructional treatment, the experimental group outperformed the control group in accu-
rately using the target form in spontaneous speech (Lindseth, 2016). However, Lindseth
(2016) did not administer any test of explicit or implicit knowledge of the target form, and
it is also not clear whether any other type of instruction would lead to a greater gain than
 Language Awareness 5

the output based activities. Therefore, the findings of previous studies investigating the
development of L2 knowledge as a function of being exposed to different types of L2 instruc-
tion have been largely inconclusive, which calls for more research in this area.

Rationale and research questions

In time-pressured tests such as a timed grammaticality judgement test (TGJT), EIT, or an oral
narrative test that were used as measures of implicit knowledge in previous studies (e.g. R.
Ellis, 2005), time pressure could only make access to explicit knowledge difficult but not
impossible (Suzuki & DeKeyser, 2015). Additionally, as previous studies on PI and OBI mostly
used ‘controlled or non-spontaneous assessment tasks’ (e.g. sentence completion, multiple
choice), that tap into explicit knowledge (Leow, 2007, p. 43), for measuring learners’ gain,
further research is warranted to investigate whether such instructional methods also lead
to the development of automatised explicit knowledge. Furthermore, morphological and
syntactic target forms need to be included in the same study to examine whether or not the
effects of instructional treatments differ depending on target structures.
To address these gaps in the literature, the present study, adopting the theoretical orien-
tation of DeKeyser (2015), examines the development in explicit and automatised explicit
knowledge of one morphological (third person singular ‘s’) and one syntactic (future condi-
tional if-clause) target forms from two types of instructional treatments: PI and OBI. The
present study focuses on automatised explicit knowledge instead of implicit knowledge
because the study is set in an instructional context where the target forms are taught to
adult ESL learners through explicit rule instruction (DeKeyser, 2015; Jiang, 2007), and a real-
istic goal for instructed learners is not only to gain explicit knowledge of L2 forms but also
to achieve some degree of automaticity in using those forms smoothly and accurately
(DeKeyser, 2015). Thus, even during rapid and effortless use of target forms in a time-pres-
sured test (e.g. oral narrative test, EIT, or TGJT), instructed learners may still be aware of the
target form rules, which is the main distinguishing criteria between automatised explicit
knowledge and implicit knowledge (DeKeyser, 2009; Suzuki & DeKeyser, 2015, 2017a).
The study was guided by the following research questions:

1. a. Can learners develop explicit and/or automatised explicit knowledge of third person
`s’ through PI or OBI?
b. If yes, is there any difference between PI and OBI in the development of explicit and/
or automatised explicit knowledge?
2. a. Can learners develop explicit and/or automatised explicit knowledge of future condi-
tional if-clause through PI or OBI?
b. If yes, is there any difference between PI and OBI in the development of explicit and/
or automatised explicit knowledge?

Methodology
Participants
The participants included a total of 72 ESL learners who were enrolled in three intermediate
level IEP (Intensive English Programme) Structure and Composition classes at a public
6 T. MOSTAFA AND Y. KIM

university in the USA (46 female and 26 male students). Those three intact classes were
randomly assigned to two experimental groups (PI [n = 31] and OBI [n = 22]) and one control
group (n = 19). The average age of the participants was 24.7 (minimum= 17, maximum = 47,
standard deviation = 6.88). The participants were from a variety of L1 backgrounds such as
Spanish, Arabic, Chinese, Vietnamese, Turkish, Cambodian, Portuguese, Azerbaijanian,
Persian, French, and Korean. The participants’ length of stay in the USA ranged from 2 to
18 months.

Target structures
The current study targeted two English as L2 forms. The morphological target was the third
person singular ‘s’ form. In Pienemann and Johnston’s (1987) developmental stages for
English morphemes, third person singular ‘s’ marking is placed in stage 5 that requires the
prior knowledge of identifying different grammatical categories in a sentence. At the inter-
mediate level, learners are supposed to be familiar with different word classes and parts of
speech as well as forming questions by moving elements in a sentence. Hence, they should
be developmentally ready to acquire the third person singular ‘s’ structure.
Additionally, the syntactic target form was the future conditional if-clause (If + pres-
ent + future) (Celce-Murcia & Larsen-Freeman, 1999, p. 545). This is a complex structure that
consists of one dependent clause (the if-clause) that ‘sets up the condition’ (Celce-Murcia
& Larsen-Freeman, 1999, p. 546) and one main or independent clause that gives the ‘result
or outcome’ (Celce-Murcia & Larsen-Freeman, 1999, p. 546). In Izumi and Bigelow (2000),
the experimental group learners could not acquire past hypothetical conditionals from
output based instruction whereas Song and Suh (2008) found positive effects of such
instruction on explicit knowledge of the same form. Such conflicting findings point to the
need of more research into L2 learners’ acquisition of conditional sentences from instruc-
tional treatments.

Measurement of explicit and automatised explicit knowledge

In the present study, explicit knowledge was operationalised as L2 learners’ ability to identify
and correct grammatical errors and verbalise the correct grammar rules using metalan-
guage (DeKeyser & Criado, 2012). Learners’ explicit knowledge was measured in an error
correction test (Sheen, 2007). In this test, there were 16 sentences, which contained third
person singular ‘s’ error (n = 8) and if-clause error (n = 8). In each sentence, the participants
were required to identify the location of the error, correct it, and write down the correct
grammar rule, and thus, this test tapped into the participants’ conscious knowledge of the
target form rules.
On the contrary, automatised explicit knowledge was operationalized as the participants’
ability to use the target forms with minimum attentional resources and without focus on
grammatical accuracy (Jiang, 2007). Learners’ automatised explicit knowledge was measured
in a timed oral narrative production test1 in which each participant read a short story twice,
and immediately after, they had to retell the story within two minutes (Ellis, 2005). The story
for the oral narrative test was written (by the first author) in present tense including examples
of third person ‘s’ (n = 16) and future conditional if-clause (n = 7) so that a retelling of it would
 Language Awareness 7

elicit the target forms. Prior to its administration to the participants, it was pilot tested with
four ESL learners.
In order to avoid the practice effect, three parallel versions of each test (oral narrative and
error correction) were designed for the three test-administrations (pre-test, post-test, delayed
post-test).

Instructional treatment materials

PI materials
Following the principles of PI, the participants of the PI group were first presented (via a
PowerPoint presentation) with the explicit rules of third person singular ‘s’ and future con-
ditional if-clause with a few examples of each (Wong, 2004). The examples were prepared
by the first author, and they were not based on any textbook. Then they were given an
explanation on L2 learners’ tendency to process the lexical items first before processing any
grammatical markers if both encode the same meaning, as was expounded in the Lexical
Preference Principle by VanPatten (2015). Then, the learners in this group participated in 10
structured input activities in which the participants had to focus on the meaning in input,
and they did not produce any output containing the target forms (Lee & VanPatten, 1995)
(see Appendices A and C for sample structured input activities).

OBI materials
The participants in the OBI group received the same explicit instruction on the target form
rules as the PI group via a PowerPoint presentation before they participated in any activities.
But unlike the PI group, the OBI group did not receive any instruction on avoiding the wrong
processing strategy. The OBI group participated in 10 meaning based activities in which they
had to produce either spoken or written output (see Appendix B for a sample output based
activity). The activities were modelled closely on those of the PI group, but it was ensured
that the participants were not presented with any input containing the target forms and
that they were producing output.

Control group
The control group only participated in three tests and received no instructional treatment.

Procedure
The data were collected over 3 weeks. Figure 1 shows the overall procedure of data collection.
All the instructional treatments were conducted by the first author. For each experimental
group, the instructional treatment took place over two consecutive class periods. On the
first day, all three groups took the pre-test. After the pre-test (20 min), both PI and OBI groups
completed the instructional treatments. The PI group received explicit instruction of the
target form rules followed by explanation of the wrong processing strategy (seven minutes),
while the OBI group received only explicit instruction of the target form rules (seven minutes).
Then the PI group participated in two structured input activities and the OBI group did two
output based activities (23 minutes for each group). After two days, on the second treatment
8 T. MOSTAFA AND Y. KIM

Figure 1. Design of the study.

day (total time 1 hour 45 minutes), the researcher again presented the target form rules with
a review of the wrong processing strategy for the PI group and only presented the target
form rules for the OBI group (five minutes). Then, the PI and the OBI groups participated in
eight structured input activities and eight output based activities, respectively (80 minutes
for each group). Immediately after the activities, both groups took the first post-test (20 min-
utes for each group). On these two days, the total time spent on instruction and activities
(excluding the tests) for each experimental group was approximately 1 hour and 55 minutes.
After two weeks, all the groups took the delayed post-test (20 minutes).

Scoring of the tests

Oral narrative production test
Each student’s retelling of the story was transcribed manually, and from each transcrip-
tion, the total numbers of suppliance and accurate suppliance of each target form were
identified. For either target structure, there was no example of over suppliance (use of
the structures in non-obligatory contexts) in the entire dataset. Therefore, we focused
on the suppliance and accuracy rate in the current study. The participants received one
point for each suppliance of each target form and one point for each accurate suppliance.
Thus, the production score, which was the sum of the suppliance and accurate suppliance
scores of each target form, was used for statistical analysis. The first author coded all the
oral narrative data, and a second rater coded 34% of the data that was randomly selected.
 Language Awareness 9

The rate of agreement between the two raters was 98%. All the disagreements were
resolved through discussion.

Error correction test

In the error correction test, the learners received one point for identifying the error in each
sentence, one point for correcting it, and one point for writing down the correct grammar
rule. Hence, the maximum score for each item was three. Thus, the total point possible for
the third person ‘s’ structure was 8 × 3 = 24, and likewise, for the if-clause, it was 8 × 3 = 24.
The first author did all the scoring, and 40% of the tests were randomly chosen to be rated
by a second rater. Cohen’s Kappa was calculated for interrater reliability. The Kappa mea-
surements of agreement between the two raters were 0.88, 0.85, and 0.93 for the pre-test,
post-test, and delayed post-test, respectively. Any disagreement was resolved through
discussion.

Data analyses
To answer the research questions, linear mixed effect (LME) analyses were performed to
account for the random variances associated with the participants at different time periods.
The lme4 package (version 1.1-15) in the software program R (version 3.4.3) was used for
the LME analysis (R Core Team, 2015). The R function r.squared GLMM was used to calculate
the effect sizes where the marginal r squared (R2m) indicates the variance explained by the
fixed factors and the conditional r-squared (R2C) indicates the variance explained by both
the fixed and random factors.
To answer the first research question, two main LME models were created: models 1 and
2. In model 1, error correction test score of third person ‘s’ was the dependent or response
variable, and time (pre-test/post-test/delayed post-test) and group (input/output/control)
were the fixed factors or independent variables. In model 2, oral narrative production score
of third person ‘s’ was the response variable, and time and group were the fixed factors.
Similarly, to answer the second research question, two more main LME models were created:
models 3 and 4. In model 3, error correction score of if-clause was the response variable
with group and time as the fixed factors. Furthermore, in model 4, oral narrative production
score of if-clause was the response variable, and group and time were the fixed factors. In
each of these models, participant was entered as the random intercept to account for the
variability related to the participants. Additionally, as the fixed factors (group, time) each
had three levels, control group and pre-test were set as the reference in each model for the
group and the time variable, respectively. Each LME model was fit using a restricted max-
imum likelihood (REML) technique. A fixed factor was considered significant if the p-value
was below .05.

Results
One-way ANOVAs and non-parametric Kruskal-Wallis tests were conducted on the error
correction and oral narrative pre-test scores of the target forms, respectively. The results
showed no significant difference between the three groups in their pre-test scores of either
target form: if-clause error correction, F (2, 70)=1.215, p = 0.303; third person ‘s’ error
10 T. MOSTAFA AND Y. KIM

correction, F (2, 69) = 2.889, p = 0.06; if-clause oral production, p = 0.39; third person ‘s’ oral
production, p = 0.59.

Explicit and automatised explicit knowledge of third person ‘s’

The first question asked whether the learners developed explicit and/or automatised explicit
knowledge of third person ‘s’ after receiving PI or OBI. The mean (M) and standard deviations
(SD) of the error correction and oral narrative test scores of third person ‘s’ are shown in Table
1 and Table 2, respectively.
In the output of the main LME model 1 (with error correction scores of third person ‘s’ as
the response variable), as reported in Table 3, there was a significant increase in scores of PI
and OBI groups compared to the control group and also a significant increase of scores at
the post-test and delayed post-test times in comparison to the pre-test time.
The coefficients of ‘Post-test’ and ‘Delayed post-test’ indicate that from the pre-test to the
post-test and delayed post-test times, the scores significantly increased by 4.338 and 3.894,
respectively. Additionally, the coefficients of ‘Input’ and ‘Output’ indicates that from the con-
trol to the PI and OBI groups, the scores significantly increased by 3.244 and 5.046, respec-
tively. This model reports a marginal R2 of 0.205 and a conditional R2 of 0.585, indicating that

Table 1. Error correction scores of third person ‘s’.

IP (n = 31) OP (n = 22) Control (n = 19)
M SD M SD M SD
Pre-test 15.29 7.363 17.818 4.982 13.105 5.685
Post-test 20.258 4.531 21.954 3.605 16.578 5.347
Delayed post-test 19.774 4.326 20.954 4.961 16.842 6.379
Note: M = Mean, SD = Standard deviation.

Table 2. Oral narrative scores of third person ‘s’.

IP (n = 31) OP (n = 22) Control (n = 19)
M SD M SD M SD
Pre-test 7.419 7.575 5.909 4.878 4.526 3.762
Post-test 5.806 5.016 7.818 7.248 6 5.207
Delayed post-test 6.709 6.056 6.727 7.369 6.316 5.783
Note: M = Mean, SD = Standard deviation.

Table 3. The main LME model on error correction scores of third person ‘s’.
Fixed effects Random effects
Standard By participant
Coefficient Error t-value p-value SD
Intercept 12.452 1.051 11.846 <.001** 3.722
Input 3.244 1.255 2.584 .009** —
Output 5.046 1.351 3.734 <.001** —
Input versus output 1.802 1.212 1.487 .137 —
Post-test 4.338 0.649 6.678 <.001** —
Delayed post-test 3.894 0.649 5.994 <.001** —
Post-test versus delayed 0.444 0.648 0.686 .493 —
post-test
Note: SD = Standard deviation.
 Language Awareness 11

the significant fixed factors collectively explained about 20.5% variance in the error correc-
tion test scores targeting third person ‘s’.
In terms of the oral narrative scores, the output of the main model 2 with oral narrative
scores of third person ‘s’ as the response variable did not have any significant predictor as
shown in Table 4.
The first research question also asked whether there was any significant difference
between the two experimental groups in the development of explicit and automatised
explicit knowledge of third person ‘s’. To answer this question, two interaction models (one
for each test) were created with interaction between group and time as the fixed factor. In
the interaction model with the error correction test scores of third person ‘s’ as the response
variable, as reported in Table 5, there was no significant difference between the PI and the
OBI groups at either the post-test or the delayed post-test.
Additionally, in the output of the interaction model with the oral narrative production
score of third person ‘s’ as the response variable, as reported in Table 6, the coefficient of
the significant predictor ‘Input versus output × post-test’ indicates that compared to the
PI, the OBI group’s post-test score was higher by 3.522. Thus, the OBI outperformed the

Table 4. The main LME model on oral narrative production scores of third person ‘s’.
Fixed effects Random effects
Standard By participant
Coefficient Error t-value p-value SD
Intercept 5.304 1.244 4.264 <.001** 4.840
Input 1.143 1.525 0.749 .456 ­—
Output 1.316 1.642 0.801 .426 —
Input vs. Output 0.173 1.475 0.117 .907 —
Post-test 0.227 0.62 0.366 .715 —
Delayed post 0.366 0.62 0.59 .556 —
Post-test vs. delayed post −0.139 0.618 −0.225 .822 —
Note: SD = Standard deviation.

Table 5. The interaction model on error correction scores of third person ‘s’.
Fixed effects Random effects
By participant
Coefficient Standard Error t-value p-value SD
Intercept 12.753 1.229 10.375 <0.001** 3.709
Input 2.537 1.566 1.620 .107 —
Output 5.064 1.684 3.007 .003** —
Input versus output 2.528 1.505 1.680 .093 —
Post-test 3.533 1.286 2.747 .007** —
Delayed post-test 3.796 1.286 2.951 .004** —
Post-test versus delayed −0.263 1.272 −0.207 .836 —
post-test
Input × post-test 1.434 1.627 0.882 .379 —
Output × post-test 0.603 1.747 0.345 .730 —
Input versus −0.8314 1.5466 −0.538 .591 —
Output × Post-test
Input × delayed post-test 0.687 1.627 0.422 .673 —
Output × delayed post-test −0.66 1.747 −0.378 .706 —
Input vs. output × delayed −1.347 1.546 0.871 .384 —
post-test
Note: SD = Standard deviation.
12 T. MOSTAFA AND Y. KIM

PI group in the oral narrative post-test scores. This interaction model reported a marginal
R2 of 0.021 and a conditional R2 of 0.644, indicating that the significant fixed factor
explained about 2% variance in the oral narrative production scores targeting third
person ‘s’.

Explicit and automatised explicit knowledge of if-clause

The second research question asked whether the learners developed explicit and/or autom-
atised explicit knowledge of future conditional if-clause after getting PI or OBI. The mean
(M) and standard deviations (SD) of the error correction and oral narrative test scores of
if-clause are shown in Table 7 and Table 8 respectively.
In the output of the main LME model 3 with error correction scores of if-clause as the
response variable, as reported in Table 9, the coefficients of the predictors ‘Post-test’ and
‘Delayed post-test’ suggest that from the pre-test to the post-test and the delayed post-test
times, the error correction scores of if-clause significantly increased by 5.443 and 4.512,

Table 6. The interaction model on oral narrative production scores of third person ‘s’.
Fixed effects Random effects
Coefficient Standard Error t-value p-value By participant
SD
Intercept 4.526 1.376 3.289 .001** 4.845
Input 2.893 1.756 1.648 .102 —
Output 1.383 1.889 0.732 .465 —
Input versus output −1.510 1.693 −0.892 .372 —
Post-test 1.322 1.204 1.098 .274 —
Delayed post-test 1.638 1.204 1.360 .176 —
Post-test versus delayed −0.315 1.187 −0.266 .790 —
post-test
Input × post-test −2.935 1.522 −1.929 .056 —
Output × post-test 0.587 1.634 0.359 .72 —
Input × delayed post-test −2.347 1.521 −1.543 .125 —
Output × delayed post-test −0.819 1.634 −0.502 .616 —
Input versus output × post-test 3.522 1.443 2.440 .015** —
Input versus output × delayed 1.528 1.443 1.058 .29 —
post-test
Note: SD = Standard deviation.

Table 7. M and SD of error correction scores of if-clause.

IP (n = 31) OP (n = 22) Control (n = 19)
M SD M SD M SD
Pre-test 12.677 5.473 11.773 4.869 10.368 4.787
Post-test 18.226 5.339 19 5.416 13.368 6.825
Delayed post-test 17.097 4.686 17.227 6.934 13.737 6.640
Note: M = Mean, SD = Standard deviation.

Table 8. M and SD of oral narrative production scores of if-clause.

IP (n = 31) OP (n = 22) Control (n = 19)
M SD M SD M SD
Pre-test 1.548 2.142 0.954 1.253 1.473 1.541
Post-test 1.935 2.279 2.773 2.562 2.052 1.649
Delayed post-test 2.129 1.995 3 2.690 2.316 2.029
Note: M = Mean, SD = Standard deviation.
 Language Awareness 13

respectively. From the control to both the PI and the OBI groups, the scores also significantly
increased by 3.606.
Additionally, the output of the main LME model 4 (as reported in Table 10) with oral
narrative production score of if-clause as the response variable showed that from the pre-
test to the post-test and delayed post-test times, the oral narrative scores significantly
increased by 0.891 and 1.113, respectively. However, the PI and the OBI groups did not
significantly increase their scores compared to the control group.
The second research question also asked whether there was any significant difference
between the two experimental groups in the development of explicit and/or automatised
explicit knowledge of the if-clause. To answer this question, two interaction models (one for
each test) were created with the interaction between group and time as the fixed factor. In
the output of the interaction model with error correction test score of if-clause as the
response variable, as reported in Table 11, the coefficient of the significant predictor
‘Output × post-test’ indicates that compared to the control group, the OBI group’s error cor-
rection post-test score was higher by 4.071. This interaction model reported a marginal R2
of 0.213 and a conditional R2 of 0.618, indicating that the significant fixed factors explained
about 21% variance in the error correction scores targeting if-clause.
Furthermore, the output of the interaction model with the oral narrative production score
of if-clause as the response variable, as reported in Table 12, shows that compared to the PI,
the OBI group’s oral narrative score was significantly higher by 1.431 and 1.465 at the post-test
and the delayed post-test times, respectively. This interaction model reports a marginal R2 of

Table 9. The main LME model on error correction scores of if-clause.

Fixed effects Random effects
By participant
Coefficient Standard Error t-value p-value SD
Intercept 9.075 1.115 8.141 <.001** 4.046
Input 3.606 1.340 2.691 .009** —
Output 3.606 1.442 2.500 .015** —
Input versus output −0.000 1.294 0 1 —
Post-test 5.443 0.659 8.261 <.001** —
Delayed post-test 4.512 0.659 6.848 <.001** —
Post-test versus 0.931 0.657 1.417 .157 —
delayed post-test
Note: SD = Standard deviation.

Table 10. The main LME model on oral narrative production scores of if-clause.
Fixed effects Random effects
By participant
Coefficient Standard Error t-value p-value SD
Intercept 1.295 0.399 3.248 .001** 1.317
Input −0.092 0.468 −0.197 .844 —
Output 0.279 0.504 0.555 .581 —
Input versus output 0.371 0.451 0.823 .410 —
Post-test 0.891 0.272 3.270 .001** —
Delayed post-test 1.113 0.272 4.086 <.001** —
Post-test vs. delayed −0.222 0.272 −0.818 .413 —
post-test
Note: SD = Standard deviation.
14 T. MOSTAFA AND Y. KIM

Table 11. The interaction model on error correction scores of if-clause.

Fixed effects Random effects
By participant
Coefficient Standard Error t-value p-value SD
Intercept 10.154 1.281 7.924 <.001** 4.045
Input 2.523 1.633 1.545 .124 —
Output 1.619 1.756 0.922 .358 —
Input versus output −0.904 1.571 −0.576 .565 —
Post-test 3.156 1.288 2.451 .015** —
Delayed post-test 3.525 1.288 2.737 .007** —
Post-test versus delayed −0.368 1.273 −0.289 .772 —
post-test
Input × post-test 2.392 1.628 1.469 .144 —
Output × post-test 4.071 1.748 2.328 .021** —
Input versus output × post-test 1.679 1.547 1.085 .277 —
Input × delayed post-test 0.895 1.628 0.55 .583 —
Output × delayed post-test 1.93 1.748 1.104 .271 —
Input versus output × delayed 1.035 1.547 0.669 .503 —
post-test
Note: SD = standard deviation.

Table 12. The interaction model on oral narrative production scores of if-clause.
Fixed effects Random
effects
By participant
Coefficient Standard Error t-value p-value SD
Intercept 1.452 0.476 3.053 .002** 1.323
Input 0.096 0.605 0.159 .874 —
Output −0.497 0.651 −0.764 .446 —
Input versus output −0.593 0.581 −1.021 .307 —
Post-test 0.633 0.528 1.200 .232 —
Delayed post-test 0.896 0.528 1.698 .091 —
Post-test versus delayed −0.263 0.523 −0.503 .614 —
post-test
Input × posttest −0.246 0.668 −0.369 .713 —
Output × post-test 1.185 0.717 1.651 .101 —
Input versus output × post-test 1.431 0.636 2.252 .024** —
Input × delayed post-test −0.316 0.668 −0.473 .637 —
Output × delayed post-test 1.149 0.717 1.60 .111 —
Input versus output × delayed 1.465 0.636 2.305 .021** —
post-test
Note: SD = standard deviation.

0.073 and a conditional R2 of 0.446, indicating that the significant fixed factors collectively
explained about 7.3% variance in the oral narrative production scores targeting if-clause.
The findings of the study are summarised in Table 13.

Discussion
The present study investigated the effects of PI and OBI on L2 learners’ development of
explicit and automatised explicit knowledge of third person singular ‘s’ and future conditional
if-clause. Overall, the findings suggest superior gains of the PI and OBI groups compared to
the control group in developing explicit knowledge and better performance of the OBI group
compared to the PI in developing automatised explicit knowledge of both the target forms.
 Language Awareness 15

Table 13. Summary of the results.

Target form Test Results
Third person ‘s’ Oral Narrative OP > IP in the post-test
Error Correction Both IP and OP > control
Both post-test and delayed post-test > pre-test
If-clause Oral Narrative OP > IP in both the post-test and delayed post-test
Both post-test and delayed post-test > pre-test
Error Correction OP > Control in the post-test.
Both IP and OP > control
Both post-test and delayed post-test > pre-test

Development in the explicit and automatised explicit knowledge of third person ‘s’
In the error correction test of third person ‘s,’ both the PI and the OBI groups outperformed
the control group. This finding might be attributed to the instructional treatment received
by the experimental groups. Making form-meaning connections during the PI, as emphasised
in the input processing theory (VanPatten, 2015), and producing concise and appropriate
output during the OBI, as asserted in the output hypothesis (Swain, 2000), might have facil-
itated developing explicit knowledge of a low salient form such as third person singular ‘s’
(Ellis, 2006). Additionally, although skill acquisition theory states that the type of knowledge
learners develop is specific to the activity-type they engage in during practice (Shintani
et al., 2013), the PI group, who did not produce any output containing the target forms
during the instructional treatment, still performed better than the control group in the error
correction test that involved producing the target forms. Moreover, this result underscores
the findings of Farley (2004) and Collentine (1998) who also found significant development
of both PI and OBI groups in receptive and productive knowledge tests (that mainly tapped
into conscious knowledge) of Spanish subjunctive, which is also a ‘complex and difficult
structure’ to acquire (Farley, 2004, p. 147). Similarly, Ellis (2009) found that L2 learners perform
better in tests of explicit knowledge for complex structures such as third person ‘s’.
In contrast, there was no significant development in oral narrative scores for third person
‘s’ from the pre-test time to the post-test/delayed post-test times for any of the groups. These
results contrast with those of Erlam et al. (2009) who found better scores for the OBI group
in the EIT of English articles at both the post-test and the delayed post-test times compared
to the control group. The lack of any significant gains by either the OBI or the PI groups in
the oral narrative scores in the present study might emphasise the difficulty in accessing or
retrieving morphological inflections such as third person singular ‘s’ during L2 production
(Loewen & Erlam, 2009; Pienemann, 2005; Prévost & White, 2000). Third person singular ‘s’
might also be difficult to acquire because of its multiple functionality as well as its redun-
dancy (Ellis, 2006). Moreover, in the present study, the total instructional treatment lasted
for about two hours in an interval of two days, and such a short interval between instructional
treatments is called massed practice. Whereas Suzuki and DeKeyser (2017b) found that for
Japanese as L2 learners, massed practice worked better than distributed practice (long time
intervals between treatments) for acquiring procedural skills of a present progressive mor-
pheme, distributed practice was found to be superior for learning English past tense syntax
in Bird (2010) and for English vocabulary in Nakata (2012). As the effects of distributed
practice depend on the level of complexity involved in the task (Bird, 2010), distributed
practice or an overall longer period of instructional treatment might work better in
16 T. MOSTAFA AND Y. KIM

developing automatised explicit knowledge of a complex English morphological form such

as third person singular ‘s’ (Ellis, 2002).
Additionally, although neither experimental group significantly developed the oral nar-
rative scores of third person ‘s’ from the pre-test to the post-test or the delayed post-test,
the OBI group significantly outperformed the PI group in the post-test scores. The facilitative
cognitive processes promoted by output based practice such as noticing, hypothesis for-
mulation, and testing (Muranoi, 2007) might have helped the OBI group outperform the PI
group. However, there was no significant difference in the delayed post-test scores between
the two groups, and thus, the automatised explicit knowledge of third person ‘s’ as developed
by the OBI group did not sustain over time. This finding runs counter to the meta-analysis
of Shintani et al. (2013) who found that in the long term, PBI, compared to CBI, led to more
durable productive L2 knowledge that might be implicit knowledge. Hence, more empirical
investigations are needed to find a clearer answer to whether OBI can better develop autom-
atised explicit knowledge of complex morphological forms such as English third person ‘s’
compared to PI.

Development in the explicit and automatised explicit knowledge of if-clause

Regarding the development of explicit knowledge of if-clause, both the experimental groups
outperformed the control group in the error correction test with no significant difference
between the two experimental groups’ performance. Similar to the findings on third person
‘s’, this result also emphasises the importance of the instructional treatments motivated by
the input processing theory (VanPatten, 2015) and output hypothesis (Swain, 2000) in devel-
oping explicit knowledge of the if-clause. The OBI group might have had a better gain in
this regard because they significantly outperformed the control group in the post-test, and
similarly, in Song and Suh (2008), the output group also performed better than the non-out-
put control group in developing explicit knowledge of past conditional if-clause.
Additionally, in contrast to the lack of any significant difference between the OBI and
PI groups in EIT scores in Erlam et al. (2009), the OBI group in the present study outper-
formed the PI group in developing oral narrative production scores of if-clause in both
the post-test and the delayed post-test times. When learners produce output, it not only
helps them ‘notice’ the gap in their current interlanguage, but it may also engage them
in testing their hypotheses about the L2 forms that can ultimately lead to ‘construction
of linguistic knowledge’ (Swain, 2000, p. 102). Such beneficial effects of OBI might be
one explanation of the present findings that also underscore the findings of Shintani
et al. (2013) where PBI led to greater sustained productive L2 knowledge in long term
than CBI. Furthermore, it should be noted that the oral narrative test in the present study
involved oral production of stories, and hence, in this test, the OBI group might have
had an additional advantage. Toth (2006) and Morgan-Short and Bowden (2006) also
found OBI groups’ superior performances compared to the PI groups in developing
productive knowledge of Spanish morphemes and pronouns. Furthermore, if-clause
might not have been a totally new target form for the participants in the present study,
and compared to PI, OBI might better facilitate developing productive knowledge of
partially acquired forms that requires consolidation through use in speech or writing
(Shintani et al., 2013).
 Language Awareness 17

Limitations and future directions

There are several limitations to the study that need to be acknowledged. As this study was
conducted in a classroom context, issues of practicality (such as time constraints and ease
of administration) needed to be considered while designing the study and selecting the
tests for measuring L2 knowledge. For instance, because of time constraints, the number
of test items was small. Additionally, the current study implemented only one test, oral
narrative test, for measuring automatised explicit knowledge following R. Ellis (2005) who
used this test as a measure of implicit knowledge. This decision in the current study was
made based on the arguments in recent empirical studies (e.g. Suzuki & DeKeyser, 2015,
2017a; Vafaee et al., 2017) that showed time-pressured tests (such as EIT and TGJT) as mea-
suring automatised explicit knowledge rather than implicit knowledge. Future validation
studies are needed that would employ both oral narrative test and recent measures of
implicit knowledge (such as real-time comprehension tasks, Suzuki, 2017) to examine
whether those tests tap into different constructs. Furthermore, the sample sizes in the
experimental groups were also small. Moreover, the present study only included instruc-
tional practices in short intervals (massed practice). Future studies can compare the effects
of massed versus distributed practices on the development of L2 knowledge because
proper arrangement of L2 learners’ instructional practice schedules is important for autom-
atisation of L2 skills (Suzuki & DeKeyser, 2017b). Finally, similar to previous classroom based
studies (e.g. Erlam et al., 2009; Loewen & Erlam, 2009) where instructional treatments were
conducted by researchers who were not regular teachers of the participants, in the present
study, the instructional treatments were also conducted by the first author, who was not
the teacher of the classes from which the participants were recruited. Hence it might be
that the participants’ performances in the research treatments and tests were affected by
their awareness of participating in a research study or being taught by someone who is
not their regular teacher. This might affect the validity of the findings (Salkind, 2010). Future
classroom based studies should avoid such limitations in their methodology.

Conclusion
In response to a lack of previous research examining L2 learners’ acquisition of morphological
and syntactic target forms from specific instructional treatments, the present study investi-
gated the effects of PI and OBI on adult ESL learners’ development of explicit and automatised
explicit knowledge of third person singular ‘s’ and future conditional if-clause. The study
employed a pre-test–post-test design with two experimental groups and one control group.
Overall, the findings suggest superior performance of both the PI and OBI groups in devel-
oping explicit knowledge of both target forms. The OBI group also performed better in
developing automatised explicit knowledge of both the morphological and syntactic forms
compared to the PI group, although this better performance of the OBI group sustained
over time for only the syntactic target form. In pedagogical contexts, teachers can provide
learners with output production activities to develop their ability to spontaneously use
specific morphological or syntactic forms during L2 production. Learners can also be engaged
in both input based and output based activities for developing their explicit knowledge of
grammar rules.
18 T. MOSTAFA AND Y. KIM

Note
1. Although in recent research (Suzuki, 2017; Suzuki & DeKeyser, 2015), TGJT and EIT were used to
measure automatized explicit knowledge, the present study used oral narrative test because of
practicality. As the data was collected from intact classrooms, each participant did not have
access to a computer or audio-player (necessary equipment for administering a TGJT or EIT).
Additionally, although oral narrative tests along with TGJT and EIT were shown to be measures
of implicit knowledge in previous studies that adopted a factor analytic approach (Ellis, 2005;
Ellis & Loewen, 2007), the validity of such interpretations was questioned in recent studies
(Vafaee et al., 2017). Furthermore, Suzuki and DeKeyser (2015) argue that in an EIT, ‘L2 speakers
might be able to monitor their utterance or use automatized explicit knowledge’ (Suzuki &
DeKeyser, 2015, p. 865) even under time pressure because ‘they have enough time (e.g., sever-
al seconds) to access explicit knowledge’ (Suzuki & DeKeyser, 2015, p. 865; 2017a). These same
arguments could be applied to oral narrative tests in which learners are also able to resort to
automatized explicit knowledge while retelling narratives (however, to our knowledge, this
has not yet been tested empirically). Therefore, the present study uses oral narrative test as a
measure of automatized explicit knowledge.

Acknowledgements
We want to express our gratitude to all the intensive English programme teachers who
allowed us to collect data from intact classrooms (special thanks to Amanda and Diana). We
also want to thank the participants who were very cooperative during the data collection
process. Finally, we express our gratitude to the anonymous reviewers whose thoughtful
feedback during the revision process helped us substantially improve the quality of this paper.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Tamanna Mostafa is a PhD candidate of Applied Linguistics at Georgia State University. Her
research interests include second language (L2) oral performances in classroom context, individ-
ual cognitive differences among L2 learners, and application of natural language processing
tools for analysing learner-language.
Dr. YouJin Kim is an associate professor of Applied Linguistics at Georgia State University. Dr.
YouJin Kim specialises in second language acquisition (SLA), second language pedagogy, and
task-based language teaching and assessment in applied linguistics.

References
Andringa, S., & Curcic, M. (2015). How explicit knowledge affects online L2 processing: Evidence from
differential object marking acquisition. Studies in Second Language Acquisition, 37(2), 237–268.
Benati, A. (2005). The effects of processing instruction, traditional instruction and meaning-output in-
struction on the acquisition of the English past simple tense. Language Teaching Research, 9(1), 67–93.
Bird, S. (2010). Effects of distributed practice on the acquisition of second language English syntax.
Applied Psycholinguistics, 31(4), 635–650.
Celce-Murcia, M., & Larsen-Freeman, D. (1999). The grammar book: An ESL/EFL teacher’s course. Heinle
& Heinle Publishers.
 Language Awareness 19

Cintron-Valentin, M., & Ellis, N. C. (2015). Exploring the interface: Explicit focus-on-form instruction
and learned attentional biases in L2 Latin. Studies in Second Language Acquisition, 37(2), 197–235.
Collentine, J. (1998). Processing instruction and the subjunctive. Hispania, 81(3), 576–587.
DeKeyser, R. M. (2009). Cognitive-psychological processes in second language learning. In M. H. Long
& C. J. Doughty (Eds.), The handbook of language teaching (pp. 119–138). Wiley-Blackwell.
DeKeyser, R. M. (2015). Skill acquisition theory. In B. VanPatten & J. Williams (Eds.), Theories in second
language acquisition: An introduction (pp. 94–112). Routledge Taylor and Francis Group.
DeKeyser, R., & Criado, R. (2012). Automatization, skill acquisition, and practice in second language
acquisition. In C. A. Chapelle (Ed.), The encyclopedia of applied linguistics (pp. 1–8). Blackwell
Publishing.
Ellis, N. C. (2006). Selective attention and transfer phenomena in L2 acquisition: Contingency, cue
competition, salience, interference, overshadowing, blocking, and perceptual learning. Applied
Linguistics, 27(2), 164–194.
Ellis, N. C. (2007). The weak interface, consciousness, and form-focused instruction: Mind the doors. In
S. Fotos & H. Nassaji (Eds.), Form-focused instruction and teacher education: Studies in honour of Rod
Ellis (pp. 17–33). Oxford University Press.
Ellis, R. (2002). Does form-focused instruction affect the acquisition of implicit knowledge? A review
of the research. Studies in Second Language Acquisition, 24(2), 223–236.
Ellis, R. (2004). The definition and measurement ofL2 explicit knowledge. Language Learning, 54(2),
227–275.
Ellis, R. (2005). Measuring implicit and explicit knowledge of a second language: A psychometric
study. Studies in Second Language Acquisition, 27(02), 141–172.
Ellis, R. (2009). Investigating learning difficulty in terms of implicit and explicit knowledge. In R. Ellis,
S. Loewen, C. Elder, R. Erlam, J. Philp, & H. Reinders (Eds.), Implicit and explicit knowledge in second
language learning, testing, and teaching (pp. 143–166). Multilingual Matters.
Ellis, R., & Loewen, S. (2007). Confirming the operational definitions of explicit and implicit knowledge
in Ellis (2005). Studies in Second Language Acquisition, 29(01), 119–126.
Erlam, R. (2003). Evaluating the relative effectiveness of structured-input and output-based instruc-
tion in foreign language learning. Studies in Second Language Acquisition, 25(4), 559–582.
Erlam, R., Loewen, S., & Philp, J. (2009). The roles of output-based and input-based instruction in the
acquisition of L2 implicit and explicit knowledge. In R. Ellis, S. Loewen, C. Elder, R. Erlam, J. Philp, &
H. Reinders (Eds.), Implicit and explicit knowledge in second language learning, testing, and teaching
(pp. 3–25). Multilingual Matters.
Farley, A. P. (2004). The relative effects of processing instruction and meaning-based output instruc-
tion. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 143–168).
Erlbaum.
Hulstijn, J. H. (2005). Theoretical and empirical issues in the study of implicit and explicit second-lan-
guage learning: Introduction. Studies in Second Language Acquisition, 27(02), 129–140.
Izumi, S., & Bigelow, M. (2000). Does output promote noticing and second language acquisition?TESOL
Quarterly, 34(2), 239–278.
Jiang, N. (2007). Selective integration of linguistic knowledge in adult second language learning.
Language Learning, 57(1), 1–33.
Krashen, S. D. (1999). Seeking a role for grammar: A review of some recent studies. Foreign Language
Annals, 32(2), 245–257.
Lee, J. F., & Benati, A. (2007). Comparing modes of delivering processing instruction and mean-
ing-based output instruction on Italian and French subjunctive. In J. F. Lee & A. Benati (Eds.),
Delivering processing instruction in classrooms and in virtual contexts: Research and practice (pp. 99–
136). Equinox.
Lee, J. F., & VanPatten, B. (1995). Making communicative language teaching happen. McGraw-Hill.
Leow, R. P. (2007). Input in the L2 classroom: An attentional perspective on receptive practice. In R. M.
DeKeyser (Ed.), Practice in a second language: Perspectives from applied linguistics and cognitive psy-
chology (pp. 21–50). Cambridge University Press.
Lindseth, M. (2016). The effects of form-focused instruction on the acquisition of subject-verb inver-
sion in German. Foreign Language Annals, 49(1), 10–22.
20 T. MOSTAFA AND Y. KIM

Loewen, S., & Erlam, R., E. R. (2009). The incidental acquisition of third person -s as implicit and
explicit knowledge. In R. Ellis, S. Loewen, C. Elder, R. Erlam, J. Philp, & H. Reinders (Eds.), Implicit and
explicit knowledge in second language learning, testing, and teaching (pp. 3–25). Multilingual Matters.
Loewen, S., & Sato, M. (2017). Instructed second language acquisition: An overview. In S. Loewen & M.
Sato (Eds.), The Routledge handbook of instructed second language acquisition (pp. 1–20). Taylor &
Francis.
Morgan-Short, K., & Bowden, H. W. (2006). Processing instruction and meaningful output-based in-
struction: Effects on second language development. Studies in Second Language Acquisition, 28(01),
31–65.
Muranoi, H. (2007). Output practice in the L2 classroom. In R. M. DeKeyser (Ed.), Practice in a second
language: Perspectives from applied linguistics and cognitive psychology (pp. 51–84). Cambridge
University Press.
Nakata, T. (2012). Effects of expanding and equal spacing on second language vocabulary learning:
Do the amount of spacing and retention interval make a difference? Paper presented at EUROSLA
Conference, Poznan, Poland.
Pienemann, M. (2005). An introduction to processability theory. In M. Pienemann (Ed.), Cross-linguistic
aspects of processability theory (pp. 1–60). John Benjamins Publishing.
Pienemann, M., & Johnston, M. (1987). Factors influencing the development of language proficiency.
In D. Nunan (Ed.), Applying second language acquisition research (pp. 45–142). National Curriculum
Resource Center.
Prévost, P., & White, L. (2000). Missing surface inflection or impairment in second language acquisi-
tion? Evidence from tense and agreement. Second Language Research, 16(2), 103–133.
R Core Team. (2015). R: A language and environment for statistical computing. Retrieved May 15, 2018,
from https://www.r-project.org/
Reinders, H., & Ellis, R. (2009). The effects of two types of input on intake and the acquisition of implic-
it and explicit knowledge. In R. Ellis, S. Loewen, C. Elder, R. Erlam, J. Philp, & H. Reinders (Eds.),
Implicit and explicit knowledge in second language learning, testing, and teaching (pp. 3–25).
Multilingual Matters.
Sheen, Y. (2007). The effects of corrective feedback, language aptitude, and learner attitudes on the
acquisition of English articles. In A. Mackey (Ed.), Conversational interaction in second language ac-
quisition: A collection of empirical studies (pp. 301–322). Oxford University Press.
Shintani, N., Li, S., & Ellis, R. (2013). Comprehension‐based versus production‐based grammar instruc-
tion: A meta‐analysis of comparative studies. Language Learning, 63(2), 296–329.
Song, M. J., & Suh, B. R. (2008). The effects of output task types on noticing and learning of the English
past counterfactual conditional. System, 36(2), 295–312.
Suzuki, Y. (2017). Validity of new measures of implicit knowledge: Distinguishing implicit knowledge
from automatized explicit knowledge. Applied Psycholinguistics, 38(5), 1229–1261.
Suzuki, Y., & DeKeyser, R. (2015). Comparing elicited imitation and word monitoring as measures of
implicit knowledge. Language Learning, 65(4), 860–895.
Suzuki, Y., & DeKeyser, R. (2017a). The interface of explicit and implicit knowledge in a second language:
Insights from individual differences in cognitive aptitudes. Language Learning, 67(4), 747–790.
Suzuki, Y., & DeKeyser, R. (2017b). Effects of distributed practice on the proceduralization of morphol-
ogy. Language Teaching Research, 21(2), 166–188.
Toth, P. D. (2006). Processing instruction and a role for output in second language acquisition.
Language Learning, 56(2), 319–385.
Salkind, N. J. (2010). Encyclopedia of research design (Vol. 1). Sage Reference.
Swain, M. (2000). The output hypothesis and beyond: Mediating acquisition through collaborative
dialogue. In J. P. Lantolf (Ed.), Socio-cultural theory and second language learning (pp. 97–114).
Oxford University Press.
Vafaee, P., Suzuki, Y., & Kachisnke, I. (2017). Validating grammaticality judgment tests: Evidence from
two new psycholinguistic measures. Studies in Second Language Acquisition, 39(1), 59–95.
VanPatten, B. (2015). Input processing in adult SLA. In B. VanPatten & J. Williams (Eds.), Theories in
second language acquisition: An introduction (pp. 113–134). Routledge Taylor and Francis Group.
Wong, W. (2004). The nature of processing instruction. In B. VanPatten (Ed.), Processing instruction:
Theory, research, and commentary (pp. 33–63). Lawrence Erlbaum Associates, Publishers.
 Language Awareness 21

Appendix A
Sample Structured Input Activity

Put the following sentences in the correct order. Then match the sentences with the pictures
(from left to right) that show Ray’s daily routine.

a. Ray watches TV.

b. Ray walks his dog.
c. Ray goes to grocery store.
d. Ray drinks coffee in his lunch.
e. Ray eats dinner with his wife.
f. Ray reads a book before sleeping.
g. Ray wakes up in the morning.
h. Ray takes his shower.
i. Ray works at office.

Justification: This activity is based on the ‘Activity H’ in Lee and VanPatten (1995, p. 106). In this activi-
ty, the learners rearrange a set of given sentences (each featuring the target grammar form) and
match those sentences with a set of given pictures. In each of these input sentences, only the
verb-ending indicates tense because the sentences do not contain any lexical item (e.g. ‘everyday’)
that would indicate a time frame. Hence, learners are encouraged ‘to attend to the grammatical mark-
ers for tense’ (p. 100) rather than relying solely on lexical items, and they have to process the meaning
of the verb (e.g. ‘drinks’, ‘eats’) in each sentence to put them in the right order (Lee & VanPatten, 1995).
Additionally, this is an example of referentially oriented activity ‘that uses an immediate concrete
22 T. MOSTAFA AND Y. KIM

reference to ascertain the truth-value of a sentence’ (Lee & VanPatten, 1995, p. 109). The pictures are
immediate reference, and the answers can be only right or wrong. As argued by Lee and VanPatten
(1995), while doing structured input activities, learners should do referentially oriented activities first
before being engaged in more affectively oriented activities that require an opinion or a personal
response from learners.

Appendix B
Sample Output Based Activity

Look at the pictures above. This is Ray and the pictures show what Ray usually does every morning.
Write down 9 sentences (one sentence for each picture) on what Ray does every morning.

1.
2.
3.
4.
5.
6.
7.
8.
9.
 Language Awareness 23

Appendix C
Additional Structured Input Activities

Activity 1

Direction: Read the following statements. Are they true for a good student at the Intensive English
Programme (IEP)?
A good student….

True Not True

1. Gets up at 10 am.
2. Does homework.
3. Watches a lot of movies.
4. Takes regular help from tutors.
5. Does not buy any textbook.
6. Goes to library after finishing classes.
7. Spends time with friends after classes.
8. Reads English newspapers or story books
to improve English.
9. Makes friends from different countries to
practice spoken English.

Justification: This activity is based on Activity I in Lee and VanPatten (1995, p. 106). In this activity,
learners are actively engaged in attending to the input because they need to process the meaning of
the verb, containing the target third person ‘s’ form, in each sentence to respond to it as ‘True’ or ‘Not
true’. Additionally, ‘learners should focus attention during processing on the relevant grammatical
items and not on other elements of the sentence’ (Lee & VanPatten, 1995, p. 108). Hence, none of the
sentences contain adverbials such as ‘always’ or ‘regularly’ so that learners focus on verb-endings for
time frame than on other lexical items. Moreover, as argued by Lee and VanPatten (1995), since the
use of third person ‘s’ is dependent on the number and person of the subject, it might be better not
to include an explicit subject noun or pronoun in each input sentence because ‘the learner is more
likely to attend to this for person number information than to the verb ending’ (Lee & VanPatten,
1995, p. 108). Hence, in this activity, the subject is placed in the beginning, and each subsequent
sentence starts with the verb where the subject is implied.

Activity 2

Direction: Select the answer that best completes each sentence about your instructor (Diana).
Afterward, Diana will tell you if you are correct or not.

1. As soon as Diana goes home, she

a. cheques her email
b. Plays with her children
c. eats dinner
d. does something else.
2. When it is time for dinner, she
a. prepares the food
b. waits for the food
24 T. MOSTAFA AND Y. KIM

c. orders a pizza
d. helps someone to prepare the food
3. Every morning, Diana eats her breakfast and
a. reads a newspaper
b. drinks coffee
c. drives her car
d. talks over phone.
4. For her lunch, Diana
a. brings something from home
b. does not eat anything.
c. gets a sub from the student center
d. goes out with her friends.
5. Every weekend morning, Diana
a. Does exercise
b. Goes for grocery shopping
c. Makes breakfast for her kids
d. Watches news on TV
6. During holidays, Diana
a. Likes to travel
b. Takes rest at home
c. spends time with friends
d. reads books
7. In the evening, she
a. Studies for his/her classes
b. Cooks dinner
c. Spends time with family
d. Watches TV
8. Before going to bed, she
a. Brushes her teeth
b. Watches movie
c. Talks to family and friends
d. Reads a book
Justification: The activity 2 is based on the ‘Activity F’ of Lee and VanPatten (1995, p. 105) where the
learners need to choose statements that they think are true based on what they know about their
instructor. In order to indicate whether or not an option is applicable to the instructor, ‘the learner
must know what the sentence means and how the grammar encodes meaning in each’ (Lee &
VanPatten, 1995, p. 105). Hence, learners’ attention is directed to the meaning of each sentence and
to the grammatical form (third person ‘s’ marker) carrying the meaning, thus, making form-meaning
connection.

Activity 3

Direction: For each sentence in column A, indicate to which sentence in column B it is most logically
connected.
 Language Awareness 25

Column A Column B
a. If John goes to school, a. I will give him a cookie.
b. If Jim studies hard, b. He will earn a lot of money.
c. If the baby cries, c. He will live away from his family.
d. If Bill starts a business, d. He will study.
e. If Bob goes to Europe, e. He will be successful.

Justification: This activity is based on ‘Activity L’ of Lee and VanPatten (1995, p. 110). Learners need to
indicate for each clause in column A, to which clause in column B it is most logically connected (both
are input clauses). In this activity, learners need to process the verb meanings (with the target verb
forms encoding meanings) in each clause to make logical connections between them (Lee &
VanPatten, 1995).

Activity 4

Direction: Listen to the short paragraph on John’s daily routine.

Now, complete the following sentences with the information that you can remember.
a. John wakes up at _____________ in the morning.
b. John drinks three cups of ___________ in the morning.
c. If John gets time, he will read a ________________.
d. John goes to school with his _______________.
e. If John’s father is busy, he will go to school with his ____________________.

Listening Text:

John is a student at Georgia State University. He usually wakes up at 8 in the morning. But he does not like
to get up early in the morning. If he goes to school in the morning, he will drink three cups of coffee. Coffee
keeps him awake. If John gets some time, he will also read a newspaper in the morning. He likes to know
what is happening in the world. John goes to school with his father. But if John’s father is busy, he will go
to school with his mother.
Justification: This activity is based on the ‘Activity J’ in Lee and VanPatten (1995, p. 106). The learners
are exposed to both oral and written input. First, they listen to structured input, and then they receive
structured written input on verb forms while they complete the sentences with other details (not re-
lated to the target verb forms). Thus, the learners do not produce the target forms (Lee & VanPatten
1995). For completing each sentence, the learners need to recall time of day, food items, objects, and
other elements. To provide such requested information, they need to process the verb forms in each
sentence (Lee & VanPatten, 1995). For example, they need to know what ‘wakes up’ mean to recall the
required information. Thus, the purpose is to make the learners actively process the structured input.