.

l
.
.

Letirner preferences

Reid, J. (ed.). 1995. Learning Sty!es in the ESLIEFL C!assroom. New York:
Heinle & Heinle.

Oxford, R. 1990. Langnage Learning Strategies: What Every Teacher Shou!d

Know. New York: Newbury House.

LEARNER LANGUAGE

Wesche, M. B. 1981. 'Language aptitude measures in streaming, marching

studentswith methods, and diagnosis oflearningproblems' in K. Diller (ed.):

Individua! Differences and Universttfs in Lttnguage Learning Aptitude. Ro\vley,

Mass.: Newbury House. pp. 119-39.

Yorio, C. 1986. 'Consumerism in second language learning and reaching.'

Canaditm Afodern Language Review 42/3: 668-87.

In ths chapter we shift our artention away fi-om learner characreristics to the
learner's language irself. Weexamine rhe rypes oferrors rhat learners make and
discuss what their errors can tell us about their kno"vledge of the language and
rheir ability to use that knowledge. We will also look at stages and sequences
in the acquisition of particular linguisric forms, keeping in mind the role of
first language influence in second language learning.

Age ofttcquisition
Burstall, C. 1975. 'French in the primary school: The British experiment.'
Czn,1dianl11odern Language Review 3115: 388-402.
Johnson, J. and E. Newport. 1989. 'Critical period effects in second language
learning: The influence of maturational state on rhe acquisirion ofEnglish as
a second language.' Cognitive Psychology 21: 60-99.

Knowing more about the development oflearner language he1ps teachers to

assess reaching procedures in rhe light of what they can reasonably expect ro
accomplish in the classroom. As we will see, there are sorne characteristics of
learner language whch can be quite perplexing ifone does not havean overall
picture of the steps learners go through in acquiring features of rhe second
language.

Long, M. H. 1990. 'Maturadonal constraints on language development.'

Studies in Second Language Acqusition 1213: 251-85.
Patkowsl<l, IYL 1980. 'The sensitive period for rhe acquisition of syntax in a
second language.' Language Learning 3012: 449-72.

In presentingsome of the findings ofSLA research, we have included a number

of samples of learner language to illusrrare the various research findings and
ro crive you an opportuniry to pracrise analysing learner language. Of course,
tea~hers analyse learner language all the time. They try to determine whether
rheir smdents have learned what has been taught and how closely their
language marches rhe targer language. But progresscannoralways be measured
in these rerms. Somerimes movemenr from one poinr in a sequence of developmenr ro another can actually lead from apparendy correct performance
(sometimes based on rote learning or very limired knowledge) ro incorrect
performance (based on an emerging understanding of the underlying rules or
grammatical relarionships in rhe language being learned). Thus, an increase
in error may be an indication of progress. A simple example of rhis s irregular
verbs. Just like young children, second language learners usually learn the
irrecrular past tense forms ofcertain verbs befo re they learn ro apply rhe regular
sim~le past -ed marker. That means that a Ieamer who says '] buyed a bus
ticket' may know more about English grammar than one who says '! bought
a bus ticket.'

Scovel, T. 1988. A Time to Speak: A Psydmlinguistic lnquiry into the Critica!

Periodfar Humtm Speech. Cambridge, Mass.: Newbury House.
Snow, C. and M. Hoefnagel-Hohle. 1978. 'The critica! period for language
acquisition: evidence from second language learning.' Child Development
49/4: 1114-28.

The irnportance ofmttintaining the first language

Cun1n1ins, J. 1984. Bilingualisrn and Specia! Education: Issues in Assessment
and Pedagogy. Clevedon, UK: Multilingual Mattcrs.
Wong-Fillmore, L. 1991. 'When learning a second language means losing
the firsr.' Early Chi!dhood Reseanh Quarterly 613: 323-46.

64

..

72

Learner language

"'"T

Learner /anguage

The concept of learner Ianguage

In addition, ir has been observed rhar the errors are nor aI,vays 'bi-direcrional'.

A tradirional version of the Comrastive Analysis Hyporhesis (CAH) would

As we Sa\.v.in Chapter l, children do nor learn Ianguagesimplythrough imirarion

predicr rhar, where differences exisr, errors \vould be bi-direcrional, that is, for
example, French speakers Iearning English and English speakers learning
French would make errors on paraHel linguistic fearures.10 illusrrate this, let
us examine one way in \vhich French and English diffcr and ho\v rhis might
be expecred to lead to errors.

and pracnce. Instead, rhey produce sentences rhat are not like those rhev have

heard. These senrences seem to be creared on rhe basis of sorne in.rernal

proce.ss~s and kno~Iedge \vhich inreracr \vich rhe Ianguage rhey hear,
permHnng them ro d1scover rhe complexities of rhe adult lano-uao-e o-raduallv.
Child.ren's ~arly language seems besr described as a developin; sy:re~ \virh ~~
O\vn rnrenm srrucrure, not simpiy as an imperfecr imirarion of adulr
senrences.

In English, direct objecrs, \vhether nouns or pronouns, come after the verb
(for example, 'The dogears it, the dog eats the cookie'). In French, directobjecrs
\vhich are nouns follo\v the verb (for example, 'Le chien mange le bscuit'lirerally, '1.he dog ears rhe cookie'), but pronoun direct objecrs precede the
verb (for example, 'Le chien le mange'-litetaHy, 'The dog it eats'). The CAH
vvould predict rhat a narive speaker ofEnglish mighr say: 'Le chien mange le'
\vhen learning French, and rhat a narive speaker ofFrench might say 'The dog
it are' \vhen learning English.

In Chaprer l \Ve aJso sa\v rhar children's kno\vledue of rhe o-rammatical svsrem
is builr up i~ predicrable sequences. For insran~e, gram~arical morphemes
such _as rhe -1ngof the ~resent p:ogressive or rhe -ed of rhe simple pasr are nor
acqu~red ar rhe s~me time, bur in sequence. Furrhermore, rhe acquisirion of
cerr_ain grammatrcal features follows similar parterns in children in differenr
env1ronments. As children conrinue to hear and use rhcir lano-uaae rhev are
able to r:vise these sysrems in \vays \vhich gradually devel;p r~~ard~ rhe
sysrem ot an adule.

In facr;research has shown that Eng!ish speakers !earning French are more
likelyto make the predicted error than French speakers leJ.rning English. This
may be due ro rhe fact rhat English speakers learning French hear many
examples of senrences \vith subject-verb-object \.Vord order (for example, 'Le
chien mange le biscuir'). Thus they make the incorrect ;:;_ssumpticn-based
on borh rhe word order of rheir firsr language and informa-don from the
second language-that all direct objecrs come afrer rhe verb. French-speaking
learners ofEnglish, on the other hand, hearing and seeing no evidence that
English pronoun objects precede verbs, do not tend to n1ake rhis error.
Researchers have also found rhar learners have intuirions rhar certain teatures
of rheir firsr language are less !ikely to be transferable rhan orhers. For
example, most learners believe rhat idiomatic or meraphoric:d expressions
cannot si1nply be rranslared word tr \Vord.

Bur whar abour second language learning? Does ir evolve in similar wavs? Do
second language learners develop rheir O\vn Ianguage sysrem in mu~h rhe
same way as firsr language 1earners? Ho\v does insrrucrion affecr rhe lano-uao-e
acquisirion of JearnerS \Yho are exposed tO the Janauaae mainly in a f;ei;n
b
b
b
j anguage cj assroom?
l!"nril rhe late ~ 960s, mosr p~ople regarded second languagc learners' speech
si~ply as an tncorrecr vers1on of rhe targer language. Their errors \\'ere
bel1eved to be rhe result mainiy of transter from rheir firsr lan&uao-e
Contrasrive analysis \Vas the basis far idenrif~.ring differences benveen rhe fi~s~
and second language and fOr predicring areas of porenrial error. So, fOr
exar:npie, one mighr predi et rhat a speakerofFrench would be Iikelv to express
che idea. ofbe~ng cold as 'I have cold' in Engiish beca use this \Voulc be a direct
:ranslanon of rhe \vay rh~s meaning is expressed in French (j~i ftoid). And,
1ndeed, son1e errors of rhis rype do occur in Iearners' language.
As vve

SJ\V

in Chaprer 2, ho,vever, not al! errors made by second lano-uao-e

learner~ can be explained in terms of firsr language rransfer alone. A nu~b~r

of srud1es show rhar manyerrors can be explained berrer in rerms oflearners'

attempts to discover the structure of rhe language being 1earned rarher rhan
an atrempr ro transfer patrerns of rheir firsr Ianguage. Furrhermore, sorne of'
rhe errors are remarkably sin1ilar ro rhe kinds of errors made bv youna firsr
language ~earners. ~ example in Eng.Iish would be rhe use ofa regular-edpast
tense end1ng on an irregularverb (as in rhe example, 'I buyed a bus ricker').

65

A.s a result of rhe finding rhat many aspects oflearners' language could nor be
exp!ained by rhe CAH, a .number of researchers began 'ro take a diffC.renr
approach ro analysing learners' errors. This approach, \vhich developed
during rhe l 970s, became known as 'error analysis' and involved a derailed
descriprion and analysis ofrhe kinds of errors second bnguage Ic1rners make.
The goal of this research \Vas to discover whar le1rners really kno\v abour rhe
Ianguage. As Pir Corder said in a famous arricle published in 1967, when
learners produce 'corn:~cr' senrences, rhey n1ay simply be repearing somerhing
rhey have already heard; \vhen rhey produce sentences which differ from rhe
rarger language, we may assume rhar these senrences reflecr the learners'
currenr undersranding of rhe rules and parterns of rhar language. 'Error
analysis' ditTered from contrasrive analysis in rhar it dld nor set out ro predict
'
,.
"'
,.,.,.
'. ' ,.
.

73

an eff~rt to understand ho\V learners process second language data. Error

analysts was based on the assumption that, like child language, second
language learner language is a sysrem in its own righr--one which is rulegoverned and prcdicrable.

his big packet. Santa Claus no wayand no body help, so onlya way give
rhem, rhen rhree robbers ride their horse dashing through the rown.
There have saloon, theygo to drinksome beer and open the bigpackent.
They plays toys in the Bar. They meet a cow boy in rhe saloon.
(unpublished data from M. J. Martens)

Larry Selinker gave rhe name inter!dnguage to learners' developing second

language knowledge (Selinker 1972). Analysis of a learner's interlanguage
shows that lt has sorne cha.racrcristics influenced by rhe learner's previously
learned l~n~uage(~), sorne characreristics of the sccond language, and sorne
~haractensncs wh1ch seem to be very general and tend to occur in all or most
interla~guage systems. lnrerlanguages are systemaric, but rhey are also
dynan11c, continually evolving as learners receive more input and revise their
hyporheses about the second language. In the activity that follows, we will
look at son1e characreristics of interlanguage.

Manyerror types are common to both learners. Both make errors ofsubject-verb
agreemenr (for example, 'a cowboy go' and 'three robbers in che mounrain
who sees' by learner 1 and 'Sama Claus ride' and 'they plays' by learner 2).
Such errors are clearly not due ro firsr language interference. l"hey reflect
learners' understanding of rhe second language system irself rather rhan an
attempt to rransfer characrerisrics of rheir 6rst language. These are referred to
as deveiopmental errors beca use rhey are errors which might very well be made
by children acquiring English as rheir first language. Somerimes rhese are
errors of overgenera!ization, that is, errors caused by rrying to use a rule in a
conrext where it does not belong, for example, rhe -.s ending on the verb in
'they plays'. Sometimes the errors are betrer described as simplijication, where
elemenrs of a sentence are lefr out, far example, or where ali verbs have rhe
same form regardless of person, number, or tense.

Activty
The Great oy Robbery
The,following rexts were written by two learners ofEnglish, one a Frenchspeaiung secondary school student, the other a Chinese-spekingadult Iearner.
In, both cases, the learners saw a cartoon film entitled The Great Toy Robbery
(Nanonal Film Board ofCanada). Afrer v1ewrng the film, srudents were asked
to retcll the srory in wriring, as if they were telling ir to so meo ne who had noc
seen rhe film.

One can also see, especially in learner 2's text, che influence of classroom
experience. An example is the use offormulaic expressions such as 'one horse
open sleigh' which is taken verbatim from a well-known Christmas song,
which has no doubt been raught and sung in the learner's ESL class. The vivid
'dashing through the rown' probably comes from the same source.

Read the texts and examine the errots made by each learner. Do they make rhe
same kinds of errors? In what ways do the two inrerlanguages differ?

Far rhose who are familiar with ilie English spoken by native speakers of
French. sorne of the errors made by che firsr learnerwill readily be recognized
as probably based on French. Similarly, those familiar with rhe English of
Chinese speakers may recognize errors made by che Chinese learner as being
due ro the learner's attempt to use patterns of Chinese in English sentences.
These are called transfer or 'interference' errors. Ir is clear, however, thar ir is
very ofren difficulr to determine rhe source of errors. Thus error analysis has
the advanrage of permirting a description of sorne systematic aspects of
learner language, but ir does not always give us clear insights into whar causes
learners ro do whar they do. Furrhermore, as Jacquelyne Schachrer poinred
out in a 1974 arcicle, learnerssometimes avo id using cerra.in features oflanguage
which they perceive to be diflicult far them. This 'avoiliance' may lead to rhe
absence of certain errors, but it also leaves thc analysr without informarion
about rhc learners' developing inrerlanguage. That is, rhe absence of
particular features will be difficulr for che researcher or teacher to observe, but
this phenomenon of'avoidance' may also be a part of the learner's sysrematic
second language performance.

Learner 1: French1-lirst language, secondary schoo! student

During a sunny day, a cowboy go in the desert with his horse. he has a
big har. His horse eat a flour. In the same rime, Santa Clause go in a cirv
to gi:'e sorne surprises. He has a red costume and a red packet ~f
surpnses. Yo u ha ve rhree robbers in che mounrain who sees Santa Clause
with aking ofglaces that ir permirted us rosee ar a long distance. Every
robbers have a horse. They go in the way of Santa Clause, not Santa
Clause but his pocket of surprises. Afrer rhey will go in a city and they
goma saloon. [... ]
( unpublished data from P. M. Lighrbown and B. Barkrnan)

Learner 2: Chinese first language, adult

T~is year Christmas comes soon! Santa Claus ride a one horse open
;Ie1gh to sem presenr forchildren. on the backofhis body has big packet.
ir have a lor of roys. in the way he meer rhree robbers. Theywant to rake

66

76

Le11rner language

Learner !.inguage

Developmental sequences

creared as rhe accuracy score for rhis morpheme. These seores were then
ranked frorn highest to lo\vesr, giving an accuracy order for rhe morphemes.

Research on second language acquisition has reveaied rhar second lano-uage

learners, like first !anguage learners, pass rhrough sequen_ces of develop~enr~
Furrhermore, in a given language, man y of these developmenral sequences are
sirnilar for firsr and second language learners. Even among second language
learn.ers, rhese developmenra1 sequences are similar: \vhar is learned early by
one is learned early by orhers, even \vhen rhey come from differenr first
language backgrounds and differenr learning environments.

The overall results of rhe srudies suggesred an orderwhich, while not the same
as rhe developmenral sequence found in rhe first language srudies, "\V::t."i similar
amono-t;I second lanauage
learners frorn differenr first language
o
....
-backgrounds.
For example, most srudies showed a higher degree of accuracy tor plural rhan
for possessive; for -ingrhan for -ed pasr. This suggesrs rhar chis accuracy order
is nor determined entirely byrhe learners' firsr language. Ho\vever, a rhorough
review of al1 rhe 'morpheme acquisition' srudies suggesrs rhar rhe Iearners' first
language has a more imporranr influence on acquisition s~qucnces rhan sorne
researchers would claim. For example, learners whose first language has a
possessive -s form which resembles the English S(such as German) seem to
acquire this form earlier rhan rhose whose firsr Ianguage has a very different
\vay of forming rhe possessive (such as French or Spanish). There are orher
unanswered questions in rhe morpheine acquisition literature. For examp!e,
sorne of the similariries and differences observed in different sntdies seem to
be dueto the \Vay rhe language samples were collected. Neverrheless, there are
sorne very srrong patterns of similarity \vhich cannot be explained by rhe
influence of rhe firsr language alone (see Larsen-Freeman and Long 1991).

Among first language learners, rhis is perhaps nor so unexpecred beca use rheir
language Iearning is parrly ried to rheir cognirive deve!opmenr, rhar is, ro rheir
learning abour rhe relarionships among people, evenrs, and objecrs around
them. But among second language learners, \vhose experences with rhe
language may vary qure wde!y and \vhose cognirive development is
essentially stable, ir is more remarkable that developmentaI sequences are so
simlar. Furrhermore, alrhough learners obviously need ro haveopportunities
ro hear or read ccrran things before rhey begin ro use rhem, ir is not always
the case rhat rhose features of rhe Ianguage which are heard mosr frequenrly
are easiesr to learn. For example, virtually every English sentence has one or
rr1ore arricles ('a' or 'rhe'), bur many learners have grear difficulty using rhese
forms correcdy. Finally, although the Iearners' firsr language does have an
influence, many aspects of rhese developmenral srages are similar among
learners from many differenr first language backgrounds.

Negation
Anorher exarnple of rhe interacrion bet\veen developmenral sequences and
first lanauacre
influence is in the accuisirion
of negarion in Enelish.
(See
b
,
V
Schumann 1979 for a review of research on negation in second language
learning.) To a large exrent, rhe acquisition of negarive sentences by second
language learners follo ...vs a parh rhat !ooks nearly identical to the srages we sa\v
in Chapter 1 for first language acquis!tion. Wbar is differenr, ho\vever, is rhat
second language learners from different first language backgrounds behave
some\vhat differendy zuithin those stages.

In the nexr section, rhe srages of acquisirion for specific grammarical features
are presented for second language learners. In Chapter 1 we sa\v sorne
developmenral sequences for English child language acquisirion ofgrammarical
morpheme~, negarion, and quesrions. Researchers insecond languageacquisirion
have exan11ned son1e of rhese same fearures, as \vell as others.

Grammatical morphemes
Several srudies ro examine rhe deveiopmenr of grammarical morphemes have
?een carried out ~ith lear~ers \vho ha ve learned English as a second language
in a natural (non-1nstrucnonal) environmenr. These srudies \Vere done \virh
learn_:rs ofdifferenr ages and ti-om ditTerenr first language backgrounds. Like
the firsr language researchers, rhe second language researchers Iookd ar
learners' use ofgrammarical morphemes such as plural, -ing, past tense, etc.
Theyrookspeech samples from a large number oflearnersar one poinr in time
and seo red each morpheme foraccuracy in rhe learners' speech. This\vas done
by idenrit)~ing every obligarory conrext for each morpheme and dividing the
:ium?er of correcdy suppled morphemes by rhe rotal number rhar should

Stage 1
The negarive element ( usua1ly 'no' or 'nor') is typically p!::tced befo re rhe verb
or the element beng negared. Often, ir occurs as the first \vord in rhe
utterance because the subject of rhe senrence is not rhere.

'

No bicycle. No have any sand. I not like it.

67

'No' is preferred by most learners in rhis early stage, perhaps because it is the
negative tOrm rhat is easiest ro hear and recognize in rhe speech rhey are
exposed to. Iralian and Spanish speakers may prefer 'no' beca use ir corresponds
ro rhe negative form in Italian and Spanish.

77

Stage 2

Table 4. 1: Devefopmental stages far question farmation (adapted from

Ar rhis stage, 'no' and 'not' may :alternare wirh 'don'r'. However, 'don'r' is nor
marked for person, number, or tense and it may even be used befare modals
like 'can' and 'should':

Pienemann, Johnston, and Brindley 1988)

Stage 1

He don'r like ir. l don'r can sing.

Stage 2

Stage 3

Single words, formulae

'Four children?'

or sentence fragments

'A dog?'

Decla.rative word arder

no inversion, no fronting:

'lt's a n1onster in che righc corner?'

'The boys chrow the shoes?'

Learners begin ro place the negative elemenr after auxiliaryverbs like 'are', 'is',
and 'can'. But ar this srage, rhe 'don'r' form is srill nor fuHy analyscd:

Srage 3

You can nor go there. He was nor happy. She don'r !ike rice.

Stage4
'Do' is marked fortense, person, andnurnber, andmost interlanguagesentences
appear to be jusr like rhose of the rarget language:
lt doesn'r work. We didn'r have supper.

Stage 4

Froncing:
wh-fronting, no inversion:

'Where the lirtle children are?'

'What the dog are playing?'

do-fronting:

'Do you have a shoes on your picture~'

'Does in this pictue there is four
astronauts?'

other-froncing:

'Is the pcture has GVo planets on top?'

Inversion in wh- +copula and 'yes/no' questions

For sorne rime, however, learners may continue to mark tense, person, and

wh- -'...-copula:

'Where is the sun?'

number on borh rhe auxiliary and rhe verb:

auxiliary other than 'do' in

'yes/no' quesrions:

'Is there a fish in che \vater?'

I didn't wenr there. She doesn'rwants to go.

Srage 5

This sequence of stages is descriptive of the second language developmenr of

most second language learners. However, although it is true that virrually all
learners ofEnglish seem ro pass rhrough a srage offorming negacive senrences
by placing 'no' before rhe verb, sorne leamers may sray longer in rhar stage
than others. If a learner's native language forrns the negative in just rhat way
(for cxomple, Spanish 'No tienen muchos libros', 'No have many books'), ir
may take longer for rhe Iearner to notice thar native speakers ofEnglish do not
form rhe negative in rhar way. Similarly, once German speakers reach srage 3
and begin ro place rhe negative marker after rhe auxiliary, rhey may also
sometimes place ir afrer lexical verbs (for examp!e, German 'Sie kommen
nicht nach Hause', They come not home').

Scage 6

lnversion in wh- questions

inverted wh-quescions wich
'do':

'Howdoyousay[proche]?'

inverted wh- questions wirh

auxiliaries orher than 'do':

'Whar's che hoy doing?'

Complex questions
quesrion tag:
negarive question:
embedded question:

'Ic's betrer, isn'c it?'

'Whycan'tyou go?'
'Can you cell me what the date is today?'

1
1

Questions
!v1anfred Picnemann and his .coHeagues have developed a framework for
describing secon<l language question stages for learners of English from a
variery of firsr language backgrounds (Pienemann, Johnsron, and Brindley
1988). An adaptcd version of rhe srages is shown in Table 4.1.

Ir is clear &om rhis figure rhatsecond language learners learn to form quesrions
in a sequence ofdevelopment which is similar in most respecrs to first language
quesrion developmenr (seeChaprer 1). Evenlearnerswhose firsr language has
subjecr-auxiliary inversion for quesrions go r:hrough srages of using declarative
word order and a period of 'froncing' in forming quesrions in English. This
partern has also been observed in the ac~uisition ofFrench and German.
The developmental sequence for questions, while verysimilar across learners,
also appears ro be affected by first language influence. For example, even
rhough German requires subjecr-verb inverson ro form quesrions ('Knnen
sie ranzen?', 'Can they dance?'), German learners ofEnglsh will pass through

68

80

Learner !anl;uage

Learner !anguage

a phase of asking questions wirhour inversion. However, once rhey reach

srage 4 and ask English quesrions wirh subjecr-auxiJiary inversion, rhey may
assume rhar subject-verb inversion is also possible. Thus, alongside correcr

Based on the information in Table 4.1, can you identify which suge of second
language question deve!opment each question fits into?

quesrions such as 'Can I play?' one may hear quesrions such as 'Play you
baseball?'

Learner 1
1 Does a dog is black and white?
2 Where the dog is?
3 Do es the boy throw a ball r

Activity

4 How many spot the dog has?

5 lt is five questions?

Learners' questions
The quesrions in rhe charr on page 81 \vere asked by srudenrs in a grade 5
imensive ESL class in Quebec, Canada. The children (aged 10-12) are all
French-speaking and have lirde con raer \virh English ourside rheir English

Learner2
6 Do you see a dog?
7 Do the dog has a shoe?
8 The boy throw a ball ora shoe?
9 The ball is on the air?
10 The dog has a Httle spot black!

class. In their English classes they spend mosr of their time in communicative
acriviries, and rheir teachers rarely correcr rheir errors or focus on specific
poinrs of grammar. In manyways, rhese srudents have an experience of rheir
second language which is similar ro rhar of Iearners in an informal language
Iearning serring.

learner 3
11 What is the dog doing?
12
13
14
15

These questions were recorded while rhe children \vere playing a picrure
idenrificarion game. Their interlocutor \Vas looking ar a picture \Vhich was a
duplicare of one of rhe four pictures \vhich rhe studenrs cou!d see. The
children asked rhese questions in arder ro garher informarion \Vhich would
permir rhem ro guess which picrure rhe interlocutor was holding.

2
2
2
2
2

2
2
2
2

3
3
3
3
3
3
3
3
3

4
4
4

4
4

4
4
4
4
4

3
3
3
3
3

4
4
4
4
4

5
5
5
5
5

5
5

5
5
5
5

5
5

5
5

6
6
6
6

6
6
6
6
6
6
6
6
6
6

Answerkey

Leamer /: Questions I, 2, and 4 are stage 3: 'does' and 'where' and 'how many'
appear simply to be 'fronted' to forma question. Question 5 is stage 2: the~e.has
been no adjustment ro rhe word arder of a dedarative senrence; only the nsing
intonation identifies the sentence as a question. Question 3 is a b!t trlcky. !t looks
Jike a correct question, but it m;;i.y be 'correct far the wrong reason'. The evidence
from chis learner's other questions suggests that 'does' is just the forrn that is
placed in front of a sentence to make a question. That woutd rnake Que_stlon 3 a
stage 3 question, just like Question ! . lf the learner had used other fo~ms of 'do' or
other auxHiaries to form yes/no questions, it would be a stage 4 quest1on.

Activity
More about questions
Anocher group of French-speaking learners from rhe same learning context
described in rhe acrivity above \Vere asked ro judge \Vhether sorne quesrions
\.Vere correcr or noL Mosr of rhese learners produced srage 2 and 3 questions
\vhen rhey participared in rhe oral quesrions game shown. above.
The rask \vas a 'preference rask' in \vhich learners\vere presenred \virh pairs of
quesrions and asked ro judge wherher only one or rhe orher was correcr or
vvherher they were borh correcr or borh incorrect. Theyalso had rhe oprion of
saying 'I don'r know.'

Sorne of the questions rhe srudenrs judged are shown in rhe charr ,pn page 82.
Deterrriine rhe developmenra1 srage corresponding ro each quesrion and
wherher rhe quesrion is correcr or not. Remernber, sorne stage 3 questions are
actually grammarica11y correcr quesrions. Then, decide which quesrions you
think rhese learners, who produced mosdy stage 2 and 3 quesrions, \vere
\villing to accept and which rhey rejecred.

Are the children running?

Is the shoe on the grass!
How many spots does the dog have?
Did the dog catch the shoe?

Stage
1
2
2
2
2
2

Learner 2: Questions 6 and 7 are stage 3: here, 'do' seems to have been placed at
the front of the sentence. Questions 8, 9, and ! O are stage 2.
Learner 3: Questions ! 1 and 14 are stage 5: a wh-question w!Lh both inversion of
the subject and rhe auxiliary and the second verb ("doing') p!aced correcdy after
the subject. Questions ! 2, 13, and ! 5 are stage 4: correct subject-verb lnversion in
'yes/no' questlons. !t is dear that these questions are different f;-om the 'does'
questions asked by Learner 1 because there are severa! different a.uxl!iary verbs in
the 'yes/no' questions.
P hotocopable Oxford Universicy Press

69

81

Stage
2, 3, 4, 5

Question

Further analysis suggests thac the students have begun ro recognize and even
use the rule rhar requires inversion of the subject and auxiliaryverb in English
questions. This rule is similar co rhe rule for question formation in French,
their first language. However, they seem ro be transferring from French a
restrict:ion on this rule. In French, pronoun subjects but nor noun subjects
can be moved to the pasr-verbal pcsition.1'hus, the French equivalents of the
oddnumberedquestions would not be grammatical in French, and rhesrudents
rejected them in English. 1-he equivalents of the even-numbered questions
and quesrions 11-15 would be considered acceptable in French - although
rhe question formula 'esr-ce que' or inversion with an inserted prono un might
also be added to rhose in quesrions 11-15.

Correct/ Accepted/
incorrect rejected

1 \/Vhy do children !ike McDona!d's?

2 Are you a good student?

3 Are the students watching TV?
4 Can l take the dog outside?

5 Can the children speak Spanish?

6 What can we watch on TV
tonight?
7 What is your brother doing!

This aspect of the acquisition of questions is another exarnple ofhow learners'

first language interacts wirh developmenral sequencesand is discussed further
in Chaprer 6.

8 When are you going to eat

breakfast!

9 Do the teachers like to cook?

10 Do they lke pepperoni pizza?

Relative clauses

11 The teachers like to cook?

A number of srudies have found that second language learners first acquire
relative clauseswhich refer to nouns in the subjecr and direct object positions,
and only later (and in sorne cases, never) learn to use them ro modify nouns
in other sentence roles (for example, indirect object and objecr ofpreposition).
Asummary of the observed pattern of acquistion for relarive c1auses is shown
in Table 4.2. lt is referrcd ro as rhe 'accessibiliry hierarchy' and ir reflects the
apparent case with which learners have 'access' ro certain strucrures in rhe

12 The chHdren can speak Spanish?

13 'vVhy fish can live in water?

14 What your brother is doing?

15 Why chHdren like McDonald's?

Answerkey
Questions 2, 4, 6, 8, ! O are correct. at stages 4, 4, 5, 5, and 3 respectively. The
students overwhe!mingly accepted these.
-

target language.

Table 4.2: Accessibi!ity hierarchy far relative clames in English (adapted from

Questions 1, 3, 5, 7, 9 are also correct, at stages 5, 4, 5, 5, and 3 respectively. The

student overwhehningly rejeaed these.

Ooughry 1991)

Questions l ! , l 2. 13, l 4, and 15 are incorrect questions, at stages 2, 2, 3, 3, and 3

respectively. Students overwhelrning!y accepted these.

Pan of speech

Relative clause

Subject

The girl who \vas sick went home.

Drect object

The story that I read was long.

Indirect object

The man who(m] I gave rhe present to wasabsent.

Ob.iect of preposition

I found the book that John was ralking abour.

Possessive

I know the woman whose father is vsiting.

Object of comparison

The person that Susan is taller than is Mary.

Photocopiab!e Oxford Universicy Press

What can the rcsults of this preference task mean? lf rhey produced most!y
scage 2 and 3 oral questions, why <lid srudents accepr sorne srage 3 questions
and rejecr orhers? Why did rhey acceptsome srage4 and 5 questions and rejecr
orhers?
;

One possible answer ro rhcse questions lies in the subject of each senrence.
Undedine the subject of each question ('childrcn' in question l; 'you' in
qucst1on 2, etc.). Whar do younoticc?Thc correctquestionswhich thesrudenrs
acc~pte<l hav~ a pronoun subject (you, 1, \Ve, rhey). The correct questions
wh1ch rhey reJecred have a noun subject (children, fish, etc.). The incorrect
questions which rhey accepted also have noun subjei.::rs.

Unlike rhe study of grammatical morphemes, negation, and quesrions, the

study of relative clauses has not been principally inspired by research on child
langu~o-e. The hierarchy was first described in a study of languages of the
world. Edward Kcenan and Bernard Comrie (1977) found rhar languages
which included the structures at rhe botrom ofhis lisr \Vould also have those

70

84

Learner !anguage

Learner language

srares and acrivities which may las~ tOr extended periods wirhour a clear
end-point.

ar the top, bur rhe opposire was nor necessarily true. Research on rhis aspect
of second Ianguage developmenr has shown rhar if a learner can use one of rhe
strucrures ar rhe bottom of rhe lisr, he or she will probably be able to use any
rhat precede ir. On rhe orher hand, a learnerwho can produce sentences wirh
relarive clauses in rhe subjecr or direct objecr posirions will nor necessarily be
able to use relatives in anyorher posirion.

Movement through developmental sequences

We have seen in rhis secrion rhar in second Ianguage acquisirion there are
systemaric and predictable srages, or sequences, of acquisition. We have seen
examples of rhis in the developmenr of grammatic:il m0rphemes, negarives,
questions, relarive clauses, and reference to past. It is importanr ro emphasize,
however, rhar developmenral srages are not like closed rooms< Learners do not
Ieave one behind \vhen rhev enrer another. In examining a language sample
from an individual !earner, ~ne shou!d not expecr ro find a!l and only examples
of behaviours from one srage. On the contrary, at a given poinr in rime,
learners may use senrences rypical of severa! different stages. Ir is perhaps
better ro rhinkofasrageas being characrerized by rhe emergen ce and increasing
frequency of a particular form rather rhan by rhe disappear1nce of an earlier
one. Even when a more advan::ed stage comes to dominare in a learner's
speech, conditons ofstress or complexiry in a communicarive interacrion can
cause the learner to 'slip back' to an earlier stage.

Reference to past
Anocher rype ofdevelopmencal sequence has also been described. In this case,
the sequence reflects learners' changing abiliry co express the same meaning.
One example of this is the developmenr of reference ro past events. Adolescenr
and adulr learners often have importanr rhings to say abour pasr evenrs, but
their kno\vledgeo(the rargerlanguageimits their abiiry to do this. Anumber
of researchers, observing learners from different first language backgrounds
and acquiring a variery of second languages, have observed a pattern which is
similar across learners.

In rhe beginning, learners with very limired language may simply refer ro
events in rhe order in which rhey occurred or menrion a time ar place to show
rhat rhe evenr occurred in the past.

My son come. He work in restaurant.

January. lr's very cold.
Vier Nam. \Ve work roo hard.

New ways of looki.ng at first language influence

Researchers rejecred rhe interpretarion of contrastive analysis \vhich made
'transfer' or 'interference' che explanation for all of a learner's difficulties \vrh
rhe targer language. This was no doubr due in part to the facr t:har conrrasrive
analvsis \vas closelv :issoci:ired \virh behaviourisrvie\vs ofb.nguage acquisirion.
In r~jecting beh:I\;iourism, so me researchers al so discarded conrrasrive analysis
as a source of valuable information about learners' language.

Later, learners start to artach a grammarical morpheme which shows that rhe
verb is marked tOr rhe past.
The people worked in the fields.
Even aft:er t:hey begin marking p~lSc rense on verbs, however, learners may sriil
make errors such as rhe overgeneralizat:ion of rhe regular -ed ending.

There is no doubt: in che minds of most researchers and teachers, ho\vever,

thar learners draw on t:heir kno\vledge of orher b. nguo.ges as they rry to
discover rhecomplexities of rhe ne\V language Eheyare learning. We have seen
someways in which thefirsrlanguage inreracts \vith deve!opmenral sequences,
When learners reach a certain stage and perceive a similarity to [her firsr
lanauaae
t:hev rnaylinaer
lono-erarthat
srage (for examn,,le, the Spanish speaker's
00'
0
b
necration) or add a subsrage (for example, the German spcaker's inversion of
subject and lexical verbs n quesrions) to the sequencc \vhich, overall, is very
similar across Iearners, regardless of their first Ianguage. They may learn a
second language rule but restrict its app!icarion (fnr e'G1mple, the French
speaker's rejection of subject-auxiliary inversion -..vith noun subjects).

She rided her bicycle.

Anorher aspecr oflearning h\V ro refer to the past has been shown in studies
by Karhleen Bardovi-Harlig and her colleagues. They found thar learners are
more Iikely to mark pasr tense on sorne verbs than on others. For examp!e,
learners seem to recognize che need ro mark pasr tense more easily in sent:ences
s,pch as 'I broke rhe vase' and 'My sisrer fixed it with glue' t:han in senrences
such as 'She seemed happy last week' or 'My farher belonged to a club'
(Bardovi-Harlig and Reynolds 1995).
Bardovi-Harlig has suggested rhar rhese differences are due to the kinds of
meanings expressed by rhe dit1erent: verbs. Learners seem ro find ir easier to
mark past tense \vhen referring to completed evenrs t:han \Vhen referring to

71

The first lancruage rnav influence learners' inrerL:lnguage in orher\vays as well.

Th,.

,...,h,."'"'~"' ... ""..,

,..:f_'

'.....~:rl ... .-.~~ '"!-..:~!-..

T.. ,.~

~1--~

C'-L---L--- _ ____ :L-.-J

85

make rhe SAVO error, and French-speaking learners of English make rhe
SVAOerror.

appeared to be caused at least in part by learners' perceprion rhar a fearure in

the rarger language was so disrant and different from ther first language rhar
they preferred notro try ir (Schachter 1974)1

Current views of first language inRuence emphasize that there is an important

interaction involvingthe first language (or other previously learned languages),
sorne universal knowledge or processes, and the samples of rhe rarger language
which learners encounter in rhe input. In Chaprer 6, \Ve will look at ho\v
instrucrion and metalinguistic information may also contribute to this
inreracrion.

Other researchers have also found evidence oflearners' sensiriviry ro degrees

of distance or difference and a reluctance to an:empt a rransfer over roo great
a distan ce. In one very revealing smdy, Hlkan Ringbom ( 1986) found that
the 'inrerference' errors made in English by both Finnish-Swedish and
Svvcdish-Finnish bilinguals \.Vere most often traceable to Swedish, nor
Finnish. The fact that Swedish and English are closely related languages
which actualiy do share man y characterisrics seems ro have led learners to rake
a chance thar a word or a senrence srructure thar worked in S\.vedish would
have an English equivalent. Finnish, on rhe orher hand, belongs to a
comp!etdy differcnt language family. This knowledge led learners to avoid
using Finnish as aso urce ofpossible transfer, "vvhether rher own first language
was Swedish or Finnish.

Table 4.3: Adverb placement in French and English

S = Subject

V= Verb

=bject

A=Adverb

ASVO
Ofren, Mary drinks tea.
Souvent, Marie boit du th.

SVOA
Mary drinks tea often.

The risk-taking associared with this perception of similariry has its limits,
however. As we nored earlier, learners seem to know rhar idiomaric or
metaphorical uses of words are ofren quite unique ro a particular language.
Eric Kellerman (1986) found that many Dutch learners of English were
reluctant to accept certain idiomatic expressions or unusua1 uses of words
such as 'Thc: wave broke on the shore' but accepted 'He broke rhe cup' even
though both are srraighrforward translations of sentences wirh the Dutch
verb breken.

Marie boit du th souvent.

SAVO

Mary often drinks tea.

*Marie souvent boit du th.
SVAO
*Mary drinks often tea.
Marie boit souvent du rh.

Another way in which learners' first language can affect second language
acquisition is in making ir difficult for them ro no ti ce that something they are
saying is absent from rhe language as ir is used by more proficient speakers.
Lydia White ( 1989) drew attention to rhe difficulties learners may have when
sorne fearure of rheir inrerlanguage and their first language are based on
patterns which are very similar bur nor idenrical. When the learner's
interlanguage form does not cause any difficulry in communicating meaning,
the learner may find it difficult to ger rid ofit. Lydia Whire gives the example
of the restrictions on adverb placement in French and English. French and
English share considerable flexibiEty in where adverbs can be placed in simple
senrcnces (.see further discussion and references in Chapter 6). However, as
the examples in Table 4.3 show, there are sorne differences. English, but not
French, allows SAVO order; French, but nor English, allows SVAO.

Note: The asterisk (~) means ch;ir the sentence is uot grammaricaL

Summary
The focus in this chapter has been on second language acquisirion by people
who, alrhough rhey may receive sorne instruction, also have consi~erable
exposure ro ther second language in natural setrings-at work, in the
schoolyard, in rhe supermarket, or rhe neighbourhood laundromat: In gene~al,
researchers have found rhat learners who receive grammar-based 1nstrucnon
srill pass rhrough the same developmental sequences and make the same types
0 f errors as rhose \vho acquire l1guage in natural settings. For example, in
sorne of the most exrensive work on acquisirion sequen ces, Jrgen Meisel and
his colleagues Manfred Pienemann and Harald Clahsen found veryconsistent
patterns in rhe acquisirionofGetman byspeakers ofseveral Romancelang~ges
who had lirtle or no instruction in German as a second language (Me1sel,
Clahsen, and Pienemann 1981 ). Pienemann la ter found very similar patterns
in rhe acquisirion of German \.vord order by speakers of English ~vhose on~y
exposure to che language \vas in rheir universiry German classes in Australia

Second language learners have dfficulty in boch direcrions. It seems fairly

easy for French-speaking learners of English ro norice the new form and to
add SVAO to thcir repenoite and for English-speaking learners ofFrench to
add SAYO, but they ha ve far grearer difficulry getting rid of the form which
does not occur in rhe target language. English-speaking learners of French

72

,
1
1

88

Learner language

Learner language
(Pienemann 1989). In Chaprer 6 we will focus on rhe second language
acquisirion oflearners in classroom sertings. Firsr, however, we \vill look ar che

Pienemann, M., M. Johnston, and G. Brindley. 1988. 'Consrrucring an

acquisirion-based procedure for s~~Janguage assessmenr.' Studies in
Second LanguageAcquisition 1012: 217-43.

classroom irself. In Chaprer 5, we will explore rhe many "vays in which

researchers ha ve soughr to understand rhe classroom environmenr for second
language acquisirion.

Schumann, J. 1979. "The acquisirion of Englsh negJtion by speakers of

Spanish: a review of rhe lirerarure' in R. W.. i\ndersen (ed.): The Acquisition
and Use ofSpanish and Engfish as Frst and Second Langrtt1ges. Washington,
D. C.: TESOL. pp. 3-32.

Sources and suggestions far further reading

General discussion offearner fanguage

Zobl, H. 1982. 'A direcron for conrrasrive analysis; rhe compararive srudy of
developmentalsequences.' TESOL Quarterly 1612: 169-83.

Cook, V. 1991. Second Language Learning and Language Teaching. London;

Edward Arnold.

Frst language injluence

Kellerman, E. 1986. '.An eye for an eye: crosslinguisdc constrainrs on rhe
development of rhe L2 lexicon.' in E. Kellerman and M. Sharwood Smirh
(eds.): Crosslinguistic lnfluence in Second Language .Acquisition. Ne\.V York:
Pergamon, pp. 35-48.

Ellis, R. 1997. Second Language Acquisition. Oxford: Oxford Universiry

Press.
Ellis, R. 1994. The Study ofSecond Language Acquisition. Oxford; Oxford
Universiry Press.

Odlin, T. 1989. Language n-ansfir. Cambridge: Cambridge Universry Press.

Larsen-Freeman, D. and M. H. Long. 1991. An lntroduction to Second

Language Acquisition. New York: Longman.

Ringbom, H. 1986. 'Crosslinguisric influenceand the foreign language Ie~trning

process' in E. Kellerman and M. Sharwood Smirh (eds.): Crosslinguistic Inf/uence
in Second Language Acquisition. New York: Pergamon, pp. 150-72.

Lightbown, P. .i\11. 1985. 'Great expectarions: Second language acquisition

research and classroom reaching.' Applied Linguistics 612: 173-89.

Schachter,
205-14.

The concept ofinterlanguage

Corder, S. P. 1967. 'The significance oflearners' errors.' fnteniationa! Review
ofApplied Linguistics 512-3: 161-9.

J.

1974. 'An error in error analysis.' Language Leaming 2412:

Relative cluse hierarchy

Doughry, C. 1991. 'Second language insrrucrion <loes make a difference.'
Studies in Second Language Acquisition 13/4: 431-69.

Selinker, L. 1972. 'Interlanguage.' International Review ofApplied Linguistics

10/2; 209-3 l.

Eckman, F., L. Bel!, and D. Nelson. 1988. 'On rhe generalzarion of relarive
clause insrrucrion in rhe acqusirion of English as a se...cond Ianguage.' Applied
Linguistics9/I; I-20.

Develop1nent11! sequences in second language acquisition

Bardovi-Harlig, K. and D. Reynolds. 1995. 'The role oflexical aspecr in the
acguisirion of rense and aspecr.' TESOL Quarterly 2911: 107-3 l.

I-Iamilton, R. 1994. 'Is implicational generalizarion unidirection:il and

maxi1nal? Evidence from relarivizarion insrruction in a second language.'
Language Learning441I; 123-57.

lvlesel, J. M. 1986. 'Reference to pasr evenrs and actions in rhe developmenr

of na rural second language acquisirion' in C. Pfaff (ed.): First and Second
Language Acquisition Processes. Cambridge, Mass.: Ne\.vbury House.

Keenan, E. and B. Comrie. 1977. 'Noun phrase accessibiliry and Universal

Grammar.' Linguistic Inquiry 8/1: 63-99.

Mesel, J. M., H. Clahsen, and M. Pienemann. 1981. 'On derermining

developn1enral srages in natural second language acquisirion.' Studies in
Second LanguageAcquisition 312: 109-35.

Reference to pnst
Bardov-Harlig, K. and D. Reynolds. 1995. 'The role oflexical aspect in the
acquisirion of tense and aspect.' TESOL Quarterly 2911: l 07-31.

Pienemann, IvL 1989. 'Is language teachable? Psycholinguistic experiments

and hyporheses.' Applied Linguistics 10/1; 52-79.

73

89