Same Source, Different Outcomes: A Study of Swedish Influence on the Acquisition of English in Finland

Same Source, Different Outcomes: A Studyof Swedish Influence on the Acquisitionof English in Finland

Terence OdIinDepartment of Englist), Ohio State University, Coiumbus, USA

Scott JarvisDepartment of Linguistics, Ohio University. Athens, USA

With a Finnish-speaking majority and a Swedish-speaking minority, Finland offers astriking contrast in the kinds of cross-linguistic influence that can occur in the acqui-sition of English in a multilingual setting. While much previous research has lookedat the differences between Finnish and Swedish influences, our study comparesSwedish influence on the use of English by native speakers of Swedish and bynative speakers of Finnish, both positive and negative transfer being evident inthe performances. The findings indicate that although both groups show influencefrom Swedish, the patterns of transfer are by no means identical.

Keywords: language transfer, cross-linguistic influence, acquisition of English,Swedish influence

IntroductionMost people who have paid attention to research on cross-linguistic

influence are no doubt aware of problems in equating language transfer (asynonym for such influence) with the influence of one's native language onthe acquisition of a second language. For one thing, the designation of secondlanguage applies only in some settings: there are many learners in the worldwho are acquiring a third or even a fourth language. Moreover, in cases wherea third language is the target, the influence on the acquisition of a new lan-guage may originate from either the first or the second language. Needlessto say, the possible influences of a third or fourth language complicate thestudy of transfer even more. Yet whatever the complications inherent in theacquisition of languages beyond the second, research on such cases offerspromising opportunities for a better understanding of the many dimensionsof the phenomenon of cross-linguistic influence. This paper focuses on onesuch dimension: namely, the influence of one particular source language (inthis case, Swedish) on the acquisition of another language (here, English),but where the type of influence varies according to whether the sourcelanguage is the native language of the learners studied.

Cross-linguistic Infiuence in Studies of i\/lultiiingualismBefore we discuss the details of our study, it will help to review various

factors that previous research has identified as relevant to the study ofcross-linguistic influence in multilingual contexts (contexts which we define

1479-0718/04/02 124-18 $20.00/0 © 2004 T, Odlin & S. JarvisTHE INTERNATIONAL JOURNAL OE MULTILINGUALISM VoL 1, No, 2, 2004

123

124 The International Journai of Muitilinguaiism

as those involving the acquisition of any language beyond a first or a second).Although space does not allow a discussion of every possible study, we willconsider the chief factors, some of which are closely related to the results ofour own study. Research has indicated that all of the following factors interactwith cross-linguistic influence in the acquisition of languages beyond thesecond: language distance, source language proficiency, target languageproficiency, order of acquisition of the languages, activation of sourcelanguages, formality of context, and constraints on verbal memory.

Language distanceSeveral studies have now indicated that the relative similarity of two lan-

guages can promote transfer, in contrast with the weaker influence of a thirdlanguage not very similar to the other two. For example, Ahukanna et al. (1981)found English influence to be stronger in the acquisition of French than theinfluence of Igbo, the native language of the Cameroonian learners studied.Similar facilitating effects have been found in other investigations: for exampleSwedish was more influential than Finnish in the acquisition of English(Ringbom, 1987, 2001), Spanish more influential than Basque in the acquisitionof English (Cenoz, 2001), and Czech more influential than English in the acqui-sition of Russian (Duskova, 1984), Similar to what Duskova observed aboutCzech influence on Russian morphology, two recent studies (De Angelis &Selinker, 2001; Herwig, 2001) indicate that similar patterns of boundmorphology in one language lead to interlanguage forms. For instance, DeAngelis and Selinker note an Italian interlanguage verb phrase e uccido (iskilled) instead of the target Italian form e ucciso, where the use of the -ido suffixseems to result from the influence of the learner's L2, Spanish, which is muchmore closely related to the target than is the learner's Ll, English.

Source language proficiencyMultilingual individuals are rarely equally proficient in all the languages

they 'know' (a verb often used in popular descriptions of multilingual com-petence). It thus seems probable that learners will be tempted to rely on asource language they know especially well. Indeed, the results of a study bySingleton (1987) support this surmise. The learner of French that Singletonstudied was a native speaker of English but also had some knowledge of Irish,Latin and Spanish; the last of these was, however, the only language that thelearner was highly proficient in (apart from his native language), and Spanishproved to be a much stronger influence than Irish or Latin on his French inter-language, English also affected his French, but the greater similarity of Spanishto French led to results like those of the studies cited above. When learners arehighly proficient in another language, their foreign language competence mayplay a role quite different from that of the native language, as seen in a studyby WiUiams and Hammarberg (1998), The target language was Swedish, thenative language English, and German a language that the learner knew quitewell. As in the Singleton study, the learner had some knowledge of otherlanguages (French and Italian), but these had little influence. The findings

A Study of Swedish Infiuence on the Acquisition of Engiish in Finiand 125

of Williams and Hammarberg indicate that English and German playedimportant but quite different roles, the former deemed to have an 'instrumen-tal' role and the latter a 'supplier' role. Thus, the learner often relied onEnglish for task-related problems such as requesting clarification, whereasshe relied mainly on German in creating an interlanguage vocabulary.

Target ianguage proficiencyA follow-up study by Hammarberg (2001) of the same learner shows that

her patterns of phonetic transfer changed as she became more proficient inSwedish, In the earlier stages of acquisition, she relied more on German pro-nunciation patterns, whereas English patterns asserted themselves as herspeech became more fluent - despite her avowed desire not to sound English,Hammarberg attributes the change in transfer patterns to highly automaticphonetic settings of her native language that were concomitant with hergreater fluency in Swedish,

Order of acquisition of languagesIndividual learner histories can affect transfer even when a group of learners

have the same Ll, L2, and L3, as evident in research by Dewaele (1998), In thisstudy, the target language was French and the native language Dutch. Onegroup of learners had studied French as a third language (with English thesecond), while another group began their study of English after they hadstarted French, The two groups differed mainly in the relative influence ofthe source languages. The L2 French group made occasional errors influencedby English but Dutch was the chief source of transfer; in contrast, the L3French group showed many more errors influenced by their L2 English,

Activation of source languagesDewaele sees recency of use as a major factor in the results he obtained in

the study described in the preceding section. The more recently activatedthe lexicon (or other areas) of a source language, the more likely that that lan-guage will have priority over other possible contenders for influencing theinterlanguage. Although Dewaele's study emphasises activation patterns ina non-native source language (L2 English), the concept of activation can alsoapply to the influence of the native language in some contexts. In a study simi-lar to that of Ahukanna et al., (described above), Chumbow (1981) found thatCameroonians in rural areas showed relatively less influence from their L2English in L3 French in comparison with the influence from African lan-guages. City dwellers, in contrast, showed more L2 English influence in L3,and Chumbow sees this as a contrast in L2 proficiency of the urban and ruralusers of L3, Even so, it also seems likely that the rural users employed Englishless, and so such knowledge would be less activated in trying to use L3,Although urban and rural Cameroonians probably have more or less the sameproficiency in their Ll, the degree of activation of the native language seems tovary according to social circumstances.

126 The internationai Journal of Muitiiinguaiism

Formality of contextApart from residence, another social factor that can affect the extent of

transfer is the formality of the linguistic encounter, Dewaele (2001) reports thatinformal speech settings induced more cases of lexical insertions and other usesof some other language besides the target, which was French, In some cases, theintrusions came from the native language, Dutch, and in others from English,

iVIemory constraintsThe study by Williams and Hammarberg described above identified distinct

roles for the native language (English) and another source language (German)in the acquisition of the target (Swedish), As noted, German supplied the bulkof the lexical items the learner attempted to use in her Swedish interlanguage.The role of German as supplier was no doubt facilitated by the many simila-rities of vocabulary shared with Swedish, However, a second language willnot always be so similar to the third language, and in such cases, the role assupplier seems likely to be far smaller. Nevertheless, learners do sometimesinvoke the resources of a vastly different language, Schmidt and Frota(1986) note cases of Arabic lexical influence rather than Ll (English) influencein the efforts of a learner in L3 Portuguese, Selinker and Baumgartner-Cohen(1995) likewise note lexical intrusions of L2 Hebrew in the L3 German of alearner whose Ll was English, If language distance were the only relevantfactor in multilingual behaviour, such intrusions 'should not' happen, Selinkerand Baumgartner-Cohen make a plausible case for a 'talk foreign' cognitivemode assuming precedence over lexical searches of the native language insuch cases. If their surmise is correct, access to the memory resources of thenative language is at least sometimes less privileged than is access to thelexical resources of other languages previously employed in other languagecontact settings.

Same Source, Different OutcomesThe discussion in the preceding section shows that several factors can affect

the way that cross-linguistic influence will operate in multilingual encounters.Sometimes one factor seems predominant, as in the study of language distanceby Ahukanna et al.. In other cases, there may be complex interactions of two ormore factors, as in the case of Singleton's study, where language distance andsource language proficiency seem to matter. Yet, despite the varying factors atwork, a common thread runs through all the research reviewed, namely, thecomparison of the effects of different source languages. The studies of languagedistance have usually shown that cross-linguistic similarity promotes thetransfer of one language more than another when there are two markedlydifferent linguistic distances between the source and the target. The investiga-tions of proficiency factors have similarly shown one language to be a greaterinfluence than another, and likewise the studies of sociolinguistic factors, aswell as the investigation of unexpected intrusions as in the study of Selinkerand Baumgartner-Cohen,

While the many factors identified are no doubt real, another factor onemight also consider is the nature of a single source language used by different

A Study of Swedish infiuence on the Acquisition of English in Finiand 127

groups, A research question that logically follows is whether the influence ofone source language used by two different groups in a multilingual settingwill invariably have the same effects on the acquisition of a new language.Our investigation of the source language effects of Swedish hypothesises thatthe answer to the question is negative. That is, the same source language willsometimes have different effects when for one group the source is also thenative language and when for the other group the source is a non-nativelanguage,

Muitiiinguaiism in FinlandAs stated already, our study focuses on Swedish and the varying influences

it shows depending on whether or not it is the native language of learners ofEnglish in Finland, Some background on multilingualism in the Finnish con-text is thus necessary, Finland has long had a bilingual character, with about93% of the population having Finnish as their native language and another 6%having Swedish (while much smaller minorities speak Saami - also known asLapp - and other languages), Ringbom (1987, 2001) and other researchers usethe term Finn to refer to the linguistic majority and Swede to refer to the 6%minority, though it must be understood that the Swedes are also Finnishcitizens. Most Finns study Swedish in school, and Swedes similarly studyFinnish, The number of years of instruction in the second (or third) languagesvaries, and likewise the motivation and the actual achievement. For manyyears, English has been another language studied in Finland. For some stu-dents (mainly Finns), it is studied as a second language, whereas for others(mainly Swedes), it is a third language, although opposite patterns obtainfor some individuals in these two groups, Bjorklund and Suni (2000) go asfar as to claim that determining whether English is L2 or L3 in the Finnish con-text can only be done at the individual level. In any case, because of the com-plexity of the patterns of study, we will not refer to English as either L2 or L3,but simply as one of the languages in this trilingual setting.

The most studied characteristic of the acquisition of English by Finns andSwedes has been the greater success of the Swedes in comparison with theFinns, As a Germanic language similar to English in multiple ways, Swedishprovides native speakers of that language with many advantages. UnlikeSwedish, Finnish is neither Germanic nor Indo-European, and while thereare some similarities with English, the linguistic distance is great. Not surpris-ingly, Finns often make use of their knowledge of Swedish in their attempts toacquire English, with some of the transfer being positive and some negative -Ringbom and others have found many instances of the latter outcome aswhen, for example, Finns use the Swedish form bild to mean a picture, as inI have some kind of bild in my head. Finns do sometimes make comparable errorson the basis of word forms found in Finnish; yet Ringbom has noted very fewinstances. This finding resembles others not only in Finland but in other areasalso discussed in regard to language distance. Moreover, Ringbom's worksupports the notion that learner perceptions of relative language distancegreatly affect perceptions of what is and is not transferable (e,g, Ahukannaet al., 1981; Cenoz, 2001; Jordens, 1977) , While Ringbom focused on Finns'

128 The Intemational Journal of Multilingualism

interlingual identifications between Swedish and English, he acknowledgedthat Ll Swedish can lead to negative transfer; in other words, Swedes as wellas Finns may attempt to make interlingual identifications between Swedishand English that prove to be misleading.

MethodThe data for our investigation come from written descriptions of a silent

Charlie Chaplin film - narratives that were produced by 140 Finnish-speakingand 70 Swedish-speaking adolescent learners of English living in Finland whovaried in their ages and their years of language study. The film portrays ayoung woman (played by Paulette Goddard) whose circumstances duringthe Great Depression drive her to steal a loaf of bread from a bakery, afterwhich she flees from the scene and is eventually befriended by CharlieChaplin. The film segment lasts approximately 8 min, and was shown to theparticipants twice. The participants were given up to 44 min to complete theirwritten narratives, and this amount of time was generally ample (see Jarvis,1998 for a fuller description of the procedures).

Table 1 shows the primary age- and instruction-related characteristics of thesix experimental groups that took part in the study. As one can see in the table,each group consists of 35 learners who had had at least two years of priorEnglish instruction. Three of the groups of Finns were also learners of Swed-ish, and both of the groups of Swedes were learners of Finnish. The group ofFinns described in the first row of the table is different from the other three:the 35th graders of this group (F5) had studied English for over two yearsbut had never been instructed in Swedish (the different performance of thisgroup will help to show when Swedish influence does or does not occur).The other three groups of Finns had varied backgrounds in terms of theirstudy of English and Swedish. The participants in Group F7 (i.e. 7th-gradeFinnish speakers) were in their fifth year of English and first year of Swedish.The participants in Group F9A (9th-grade Finns) were in their seventh year ofEnglish and third year of Swedish, whereas the participants in Group F9B(also 9th-grade Finns) were in precisely the opposite situation: i.e. they werein their third year of English and seventh year of Swedish. The learners in

Table 1Group

F5

F7

F9A

F9B

S7

S9

Experimental participant groups

n

35

35

35

35

35

35

Ll

Finnish

Finnish

Finnish

Finnish

Swedish

Swedish

Ages

11-12

13-14

15-16

15-16

13-14

15-16

Grade

5

7

9

9

7

9

English instruction

3rd year

5th year

7th year

3rd year

3rd year

5th year

Swedish instruction

None

1st year

3rd year

7th year

Finnishinstruction

5th year

7th year

A Study of Swedish Influence on fhe Acquisifion of English in Finland 129

Group S7 (7th-grade Swedes), in turn, were in their third year of English andfifth year of Finnish, and those in Group S9 (9th-grade Swedes) were two yearsfurther along in their study of both Finnish and English. In addition to thesesix experimental groups, native control groups consisting of 66 Finns, 44Swedes, and 66 Americans were asked to produce written narratives of thefilm in their respective Lis. The relevance of these data will become clear inour presentation of the results in the following section.

ResultsIn our prior reading of the learners' narratives for other research purposes

(e.g. Jarvis & Odlin, 2000), we noted possible cases of transfer owing some-thing to Swedish influence, and accordingly we have selected four lexicalitems (from a larger number of possible cases) worth a closer look, all of themEnglish words that have Swedish cognates: instead, for, some, and what. We willfirst present the results of our analysis of the adverb instead and its preposi-tional counterpart instead of, which are similar in form, function, and distri-bution to the Swedish adverb i stdllet and preposition i sttillet for. The formalsimilarity between English and Swedish led us to wonder whether both theFinns and Swedes would readily identify instead with its Swedish equivalenti sttillet, and whether they would be less likely to make an interlingual identi-fication between instead and the corresponding Finnish adverb sen sijaan orpostposition sijaan/sijasta, whose phonetic, graphemic, and syntactic proper-ties differ substantially from those in English and Swedish, even though thediscourse functions are similar.

In our English narratives written by Finns and Swedes, we found 19occurrences of instead (sometimes written as intead or insted) produced by 19different participants. Fifteen of the occurrences were as adverbs, and theremaining four occurred as prepositions. Examples from the data are givenin Table 2, while Table 3 shows the breakdown of occurrences per languagegroup and per grammatical function for both the experimental and Ll-controldata.

Interesting differences between the Finnish and Swedish experimentalgroups can be seen in Table 3. First, only 1 (0.7%) of the 140 Finnish experi-mental participants used the word instead at all, whereas 18 (25.7%) of the

Table 2 Exannples of Swedes' use of instead

Adverb

1, They let Charlie go and took the woman instead.

2, The baker and the women run after the policeman and he took the girl instead.3, The police come and take Chaplin, but the woman tell them that is wasn't Chaplin,

it was the girl she saw it!, so they take the girl insted.

Preposition4. And then Charlie Chaplin said that it was him, and the policeman took Charlie

instead the woman.

5. Charlie is polite and stands up and she sits down instead of him.

130 The International Journal of Mulflllngualism

Table 3 Occurrence of instead (or equivalent) in the English, Finnish, and Swedish data

Finns (English)

Swedes (English)

Americans (English)

Finns (Finnish)

Swedes (Swedish)

Adverb

0

0%

15

21%

(15)

0

0%

0

0%

9

20%

(11)

Preposition

1

1%

(1)

3

4%

(3)

3

5%

(3)

0

0%

1

2%

(1)

Note: Figures in this table do not include those for group F5. The figures in the first rowof each cell are the numbers of participants who produced the respective forms, thepercentages are the proportions of participants per group who produced them, andthe parenthesised numbers are the raw occurrences.

70 Swedish experimental participants used it. The Finns' and Swedes' patternsin English also resemble their Ll patterns. Similar to the single occurrence ofinstead in the Finns' English narratives, the narratives written in Ll Finnishcontained zero instances of the corresponding adverb sen sijaan or postpositionsijaan Isijasta; and similar to the 18 occurrences of instead produced by 18(25.7%) of the Swedish experimental participants, the narratives written inLl Swedish contained 12 instances of i stallet (sometimes written as istallet)produced by 10 (22.7%) of the Swedish Ll-control participants. Besides overalluse, however, the Swedes' strong preference for using English instead as anadverb also strikingly resembles the Ll preference. However one chooses tointerpret the Finns' performance, that of the Swedes clearly indicates positivetransfer from Swedish to English.

Our second analysis concerns the preposition for, which is formally andfunctionally similar to the Swedish word for. As Table 4 shows, for and itsSwedish equivalent are both used to mark benefit and purpose, and they alsooccur in many equivalent verb-preposition collocations, such as pay for andthank for (Sw. betala for and tacka for). Unlike Swedish, the English wordcan additionally be used to indicate intended destination and duration, andunlike standard English in its modern forms, the Swedish word is usedproductively in infinitives of purpose. For and for also differ in the range ofcollocations in which they occur in their respective languages.

In contrast to Swedish, Finnish has no single close translation equivalentto the English preposition for. When comparing our Ll English and Ll Finnishdata, for example, we found that the English for was used in the same

A Study of Swedish Influence on the Acquisition of English in Finland 131

Table 4 Examples of English for and Swedish for

Benefit

1. Charlie said that I'll get a home for us.

2. Charlie steg upp for henne.'Charlie got up for her' [i.e. so that she could sit down]

Purpose

3. She would do anything for some food.4. Charlie bestammer sig for att bygga ett hus.. .aven om han maste arbeta for det.

'Charlie decides to build a house.. .even if he must work for it. '

Duration

5. So they ran again.. .and listened to heavy music for the rest of their lives.

Destination

6. They saw a man leaving for work and saying goodby to his wife.

Various collocations

7. The girl runs away but stays waiting for Chaplin.8. But he didn't pay for it.

9. Han bestammer sig for att gora sa.

'He decides to do so.'

10. En fattig flicka stal ett brod.. .men en dam sag det och berattade for bagaren.

'A poor girl stole a loaf of bread.. .but a lady saw that and told the baker.'

Infinitive of purpose11. Flickan gor sitt basta for att prata sig ur situationen.

'The girl does her best to talk herself out of the situation.

12, Charlie for in till konditoriet for att ata nagot.

'Charlie goes in to the cafe to eat something.'

functional contexts as two postpositions, one infinitival suffix, and sevendifferent nominal suffixes in Finnish. For reasons of space, we will not attemptto provide an exhaustive account of the Finnish equivalents, but we will referto some specific characteristics of Finnish in connection with our findings.

As for the results, we can generally say that the quantitative and qualitativepatterns of the use of/or by Finns and Swedes indicate that: (1) Finns who havestudied Swedish use their familiarity with for to construct novel Englishutterances with for; (2) Firms and Swedes show different pattems of the use offor in their writing, especially with regard to overgenerahsation. As mentionedin the Method section, we studied the English pattems of one group of Firms(F5) who had not studied any Swedish at all. Comparing this group with a groupof Finns who had studied both Swedish and English provides a way to get at thequestion of how some familiarity with Swedish might affect a Finn's acquisitionof Enghsh. The group most Uke F5 in terms of Enghsh study is F9B, students whohad only two years of English (but, in contrast with F5, also six years of Swedish).All instances of the use of for by both groups are listed in Table 5, with theparenthetical designations indicating the various individuals of the groups.

132 The International Journal of Mulfillngualistn

Table 5 AU uses of for by the F5 and F9B groups

F5 group

We can have a home—if I have work for it. Girl take Chaplin (F5 14)

says. I'll do it, t work much for it Chaplin says. But (F5 21)

much food, and he doesen't cost for the waitress. Chaplin (F5 42)

in to the restaurant and say for he: 'I' (F5 42)and say for he: 'I dont coas for waitress.' The policeman (F5 42)

F9B group

him. Chaplin asks a cigarette for a solder and again, he (F9B 7)

d Charlie promises to get a home for them. A police comes and (F9B 11)

The car move and stopped for a while because that girl (F9B 20)Then the girl want to go away for the car and just then the (F9B 20)

put the botle under the cow. For a while C. took it and (F9B 20)

For a while C. took it and put it for a table. They started t (F9B 20)

Charlie Chaplin. The lady told for the man who was carried (F9B 22)

the polise-man. They went looked for the girl and Charlie went (F9B 22)and policeman went looking for that girl. Charlie (F9B 23)

Charlie promised to get a house for them even if it means (F9B 25)was in the kitchen cooking a meal for them. And Charlie was (F9B 27)

house, even if I had to work for it. Then there came (F9B 27)and a girl was trying to jump away for a car but they (F9B 29)

car on the street. A girl asked for the little man 'Would you' (F9B 29)Chaplin promises to buy a house for them eaven if he (F9B 30)

eaven if he would have to work for it. Then comes a (F9B 30)

like that. We would take oranges for a window and we would (F9B 31)a cow and we would get milk for it. And we would eat (F9B 31)

home them, even he have work for it. Then the policeman (F9B 33)

He ate a lot but he didn't pay for food, so the police came (F9B 35)away. They walk along the road for quite, then they sat on (F9B 35)

kafe. He bought himself food and for girl. Then he bought (F9B 37)

caught. Charlie Chaplin said for gril that he was it who (F9B 37)

wife. She was cooking something for himself and for Chaplin (F9B 39)cooking something for himself and for Chaplin. It was some (F9B 39)

train sat man, who gave the place for girl and the talked about (F9B 45)

Although both groups have the same number of students (35), the F5 groupshows far fewer instances of/or: only three students used the word, and the totalnumber of tokens is only five. In contrast, 15 students in the F9B group used for,producing a total of 25 tokens. The evidence thus strongly suggests that previousfamiliarity with Swedish for encourages the use of English for.

A Study of Swedish Influence on fhe Acquisition of English in Finland 133

We have also found that Finns and Swedes show different patterns of theiruse of/or, especially with regard to overgeneralisation. In Table 6, the first fivecolumns represent uses of for that are appropriate in English, and some ofthese would be appropriate in Swedish as well. The next two columns,Swedish-based collocations and infinitives of purpose, represent patterns thatare appropriate for the Swedish word for but not for the English word for. Theremaining column represents overgeneralised uses of/or that are not appropri-ate in either English or Swedish.

The results in Table 6 show some similarities between the Finns and Swedesin various categories, but there are also a number of notable differences. Forexample, the Finns show a discernibly stronger tendency to use for to representbenefit than the Swedes do, and the Swedes show a clear advantage in usingEnglish-appropriate collocations. The Sw edes also use for occasionally ininfinitives of purpose, whereas the Finns never do.

Although all the differences mentioned so far can be characterised in terms ofdifferences between Ll Finnish and Ll Swedish, there are also some differencesin the Finns' and Swedes' use of for that reflect differences in the way they relyon their knowledge of Swedish. The most interesting differences related toSwedish influence involve overgeneralisations. Table 6 shows that 12 Finnsproduced 17 overgeneralisations oi for, which contrasts with the one instanceproduced by a Swede. Not all of the Finns' overgeneralisations reflect Swedishinfluence, but some most likely do. For example, most or all of the Finns' eightovergeneralisations involving verbs of saying (i.e. say for and ask for) are prob-ably influenced by the obligatory use of for in Swedish with the verb of sayingberd'tta (tell) (e.g. 'Du maste beratta for mig om din skola' is equal to 'You musttell me about your school'). Two Finns and three Swedes in the present studyused the direct Swedish-influenced collocation tell for (e.g. 'An old womansaw it and told it for the police'), and these instances are reflected in thecolumn of Table 6 that represents Swedish-based collocations.

The difference between the Finns and Swedes with respect to the use of forwith verbs of saying is that, whereas both groups occasionally used translationequivalents of Swedish collocations (i.e. tell for), only the Finns overextendedthe use of for to other verbs of saying: say for and ask for (e.g. 'Then he said forthe police, that: "I don't have money'"; 'A girl asked for the little man "Wouldyou like to come with me?'").

Ll Finnish probably did contribute to the Finns' tendency to overgeneralisethe tell for pattern to other verbs of saying (e.g. because Finnish verbs of sayingtend to share the same or similar syntactic frames), but it was almost certainlySwedish influence (i.e. the heratta for pattern) that resulted in the use of forinstead of to or other prepositional options within these overgeneralisations.To sum up, the present results suggest that Finns do rely on their knowledgeof Swedish for when using English for, but the way they rely on Swedish isdifferent from the way the Swedes rely on it.

Relativisation is another area where Swedes and Finns differ in their relianceon Swedish. In Swedish, the word som functions as a relative pronoun, as in theexample in Table 7, and it is common in the narratives of the Swedes writing intheir native language. The English word some is very similar in spelling andpronunciation, but instead of being used as a relative pronoun, it functions

134 The International Journal of Mulfilingualism

glis

h d

ata

and

fo

r in

th

e S

wed

ish

dat

aT

able

6 O

ccur

renc

e of

for

in t

he

En

Ove

rgen

eral

isat

ion

Sw

edis

h p

att

ems

En

glis

h-a

ppro

pria

te

pa

ttem

s

Infi

nit

ive

of p

urp

ose

Col

loca

tion

Col

loca

tion

Des

tin

atio

nD

ura

tion

Pu

rpos

eB

enef

itfN

O

CO

o

in

CNT—1

?:

11%

1

sso

ssCO

sso

ssin

ss1—1T-H

13%

(17)

s

(7)

g

(13)

T-H

Fin

ns(E

ngli

sh)

CO

T 1

CN

V.

T-H

s?

20%

S?CO

29%

T-H

(3)

S

T-H

CN

S

(Z)

(22)

(4)

Sw

edes

(Eng

lish

)

o

o

o

CO

CO

00

sso

sso

s?o

32%

5i

5?in

35%

27%

1

CO

(3)

(27)

(19)

Am

eric

ans

(Eng

lish

)

o

00

COT-H

o

o

Ov

T 1

sso

18%

30%

16%

sso

o

43%

CN

(8)

ONT 1

(22)

rH

Sw

edes

(Sw

edis

h)

obb

A Study of Swedish Influence on the Acquisition of English in Finland 135

Table 7 Examples of Swedish som, English some, and uses of some as a relative pronoun

The relative pronoun some

1. Efter en stund kom damen som hade stulit brodet...

'After a moment came the woman who had stolen the bread...'

The determiner some

2. After being busted he gave cigars to some kid.

The indefinite pronoun some

3. Then he eats some.

Uses by Swedish speakers of some as a relative pronoun

4. Than come the woman and said it was the girl some were took the bread.

5. .. .the man not stole the bread that were girl some stole.

primarily as a determiner or as an indefinite pronoun, as seen in the same table.Despite the S)mtactic difference, the formal similarity appears to have led threeSwedes to use English some as a relative pronoun, as in the examples given inTable 7. One of these students used it twice this way in the same paragraph,and there was a fourth student who may have considered som to be possiblein English but who marked two such uses with parentheses. This punctuationconvention could indicate either a certainty that som is not really the wordneeded, or it could at least indicate considerable doubt about the viability ofthe Swedish form in English. Nevertheless, it is hardly surprising that at leasta few Swedes would extend their interlingual identifications into the domain ofrelativisation, since the spelling and pronunciation of som and some are highlysimilar. Even so, the Finns never did, as indicated in Table 8.

Although the Finns never used some as a relative marker, the story is quitedifferent for the pronoun what. Table 9 summarises the uses of this pronounin our data. As Quirk et al. (1985) observe, nominal relatives have two nounphrase functions: What functions both as a direct object and as a subject inthe sentence The old lady sees what happen given in Table 9. In contrast, the useof what as an ordinary relative pronoun shows only one NP function, as inthe sentence But same lady what he meet come to policecar too, where what functionsas the direct object, this sentence coming from a Finn. This use of what as a sim-ple relative is, of course, an interlanguage construction which is not used in anystandard variety of the target language. As Table 10 indicates, this constructionis used only by Finns (eight of whom produced the nine cases), and that nat-urally suggests language transfer. The Finnish word mikd, which often trans-lates as what, seems to be the primary source. In our Finnish data, we found82 occurrences of mikd' in its various case forms, and six of these occurrences(produced by five different participants) showed the simple relative function.

Even though Firmish influence is primary in this case, Swedish influence on theEnglish of the Finns is, at the very least, also possible. The consonantal systems ofSwedish and Firmish are not identical, but some characteristics no doubt provideinducements for Finns to equate the Swedish vad and the English what. Forexample, there is no phonemic contrast in Finnish or Swedish between the labio-velar semiconsonant /w/ and the labiodental fricative /v/ , and although the

136 The International Journal of Multilingualism

Table 8 Occurrence ol

Finns (English)

Swedes (English)

Americans (English)

some in the English data

Determiner

36

34%

(60)

22

31%

(34)

22

33%

(35)

Indefinite pronoun1

1%

(1)

2

3%

(2)

4

6%

(4)

Relative pronoun

0

0%

4

6%

(6)

0

0%

Other

0

0%

0

0%

2

3%

(2)

Note: Figures in this table do not include those for group E5.

Table 9 Examples of what in the English data

Nominal relativepronounNominal relativedetenninerSimple relative pronounOther determinerInterrogative pronoun

The old lady sees what happen.

.. .the girl tryed to imagine what kind of life it would be...

But same lady wJwt he meet come to policecar too

Oh, but what a bad luck!

Police man askd—What happend?

articulatory contrast is not a difficult one for Firms, the spelling system does notencourage them to make the distinction because the graphemes <v> and <w>have the same pronunciation in the Finnish alphabet. Along with the phonologi-cal factors, there is another reason to believe that at least some Finns make inter-lingual identifications between Swedish vad and English what. Just as in the caseof for, there is evidence that Swedish influences the English form. In the use ofwhat, we see a difference between 105 Finns who have studied Swedish and 35Finns who have not. Table 11 compares the frequency of the use of what as a sim-ple relative by the three groups with some Swedish knowledge (F7, F9A, F9B)with the use by the group (F5) that had not had any instruction in Swedish: onlyone person in F5 used what as a simple relative. Such use of what by just one indi-vidual can be interpreted in different ways. On the one hand, it can be seen as aninfluence of Finnish operating completely independently of that of Swedish. Onthe other hand, the absence of such uses of what in the writing of 34 other 5th gra-ders not instructed in Swedish suggests that this pattern of relativisation is lesslikely when only Finnish can be drawn on as a source of hypotheses abouthow target language relative structures work. We acknowledge, however, thatthe number of tokens is small, and also that the 5th graders produced fewerinstances of targetlike as well as non-targetUke relative pronouns overall, so thisshould certainly also be taken into consideration. A similar problem worthy ofstudy in the future would be to compare whether Groups F9A and F9B show

A Study of Swedish Influence on the Acquisition of English in Finland

Table 10 Syntactic classes of English what in the film narratives

137

Finns

Swedes

Americans

NRP

27

26%

(41)

18

26%

(21)

21

32%

(26)

NRD

7

7%

(7)

0

0%

0

0%

SRP

8

8%

(9)

0

0%

0

0%

OD

3

3%

(3)

0

0%

0

0%

IP

1

1%

(1)

2

3%

(2)

0

0%

Other

4

4%

(4)

1

1%

(1)

0

0%

Note: NRP, nominal relative pronoun; NRD, nominal relative determiner; SRP, simplerelative pronoun; OD, other determiner; IP, interrogative pronoun. Figures in this tabledo not include those for group F5.

Table 11 Number of Finns using what as a simple relative pronoun in relation tonumber of years of Swedish L2 instruction

F5, n = 35(none)

1

(1)

F7, n = 35, 1styear (0.5 years)

3

(3)

F9A, n = 35, 3rdyear (2.5 years)

2

(3)

F9B, n = 35, 7thyear (6.5 years)

3

(3)

Note: Parenthetical numbers in the bottom row indicate raw tokens of use.

any difference in their use of for because of differing years of instruction inEnglish.

DiscussionWe will now review our findings and comment on the implications. We

examined the ways that Swedes and Finns show - or do not show - influencefrom Swedish in their English, our evidence coming from four words in theirwriting: instead, for, some, and what. We found striking inter-group differencesin each case. Many Swedes did use instead, while the Finns almost never did.In the case of/or, both Finns and Swedes used it, but the former tended towardmany more overgeneralisations, and Finns who had not studied Swedish veryrarely used it in their writing (in contrast to those who had studied Swedish).Finally, both Finns and Swedes used relative pronouns anomalously, but forSwedes, the pronoun thus selected was some, whereas for the Finns, it waswhat (and once again, Finns who had not studied Swedish almost never usedwhat in this way).

One methodological problem that may be evident in our work concerns theFinns' and Swedes' use of the words studied. We surmise that those studentswriting in English made interlingual identifications between Swedish andEnglish with the words in question, but the evidence for our surmise comes

138 The Intemationai Journal of Multilingualism

from the Swedish writing of other students instead of from those writing inEnglish. Ideally, we would be able to discuss the Swedish writing of the Finnsand to consider, likewise, the Swedish writing from those Swedes whoseEnglish writing we analysed. However, we simply do not have such data.On the other hand, we have offered considerable evidence suggesting the like-lihood of Swedish influence, along with the undeniable similarity of the fourwords in Swedish and English.

Three factors seem especially prominent in the results we obtained: psycho-typology, proficiency, and overgeneralisation. We will now discuss each ofthese factors and compare what we have found with results from other studiesof multilingual behaviour. Our results are consistent with those of Ringbomand other Finnish researchers with regard to psychotypology: both the Finnsand the Swedes used their knowledge of Swedish to form hypotheses aboutthe target language, English. There is a striking difference in the use of forbetween Group F9B, students who had studied Swedish for six years, andGroup F5, students who had never studied the language. Studying Swedishno doubt led to an awareness of the cognate relation between for and forand the possibility of using this relation in English. Similar results are evidentin the case of what, although the number of tokens is smaller (and there is alsoan interaction of Swedish influence with Finnish in the use of what, as the Fin-nish equivalent can function as a simple relative pronoun). These results alsoaccord well with those of Ahukanna et al. (1981) and Cenoz (2001), discussedearlier in the paper, where the similarities of two Indo-European languagesprove especially conducive to transfer.

Proficiency in Swedish also helps to explain some of the results. Asdiscussed above, a complete lack of proficiency in Swedish seems to restrictthe likely hypotheses learners will construct. Moreover, even when Finns havestudied Swedish for several years, they do not seem inclined to use somestructures that Swedes favour. The use of instead by Swedes (having, of course,native-speaker proficiency) indicates positive transfer, and the use of some as arelative pronoun indicates negative transfer; yet both structures hardly everoccur in the writing of the Finns. It is interesting to speculate on what differ-ence in results there might be if another study considers how Finns living inSweden use these two structures. Ringbom (1987, 2001) cites research indicat-ing that the second language proficiency in Swedish of Finns living in Swedenhad a facilitating effect on the acquisition of German: the longer the Finns hadlived in Sweden, the greater was the Swedish influence on their German. Suchresults are rather symmetrical with those of Hammarberg (2001), who dis-cusses the role of German as a supplier language (as summarised earlier in thispaper); the learner described relied largely on her high proficiency in Germanas opposed to other languages she had studied or her native language, English(cf. Williams & Hammarberg, 1998).

Overgeneralisation is also evident in the results we obtained regarding theuse of for by Finns, who sometimes extended a Swedish pattern to construc-tions not used by the Swedes, as in the case of say for. Such uses imply an inter-action between transfer and overgeneralisation. In error analyses about threedecades ago (e.g. Taylor, 1975), researchers frequently dichotomised transferand overgeneralisation, and a similar dichotomy between transfer and simpli-

A Study of Swedish Influence on the Acquisition of English in Finiand 139

fication was also common. Elsewhere, we have noted problems with a similardichotomy between cross-linguistic influence and simplification (Jarvis &Odlin, 2000), since interaction of the two is evident. Researchers will likewisedo well to consider that there is an interaction between cross-linguistic influ-ence and overgeneralisation. One problem with doing so, however, is that allthree terms mean different things to different researchers, and one has to lookclosely at the actual evidence in any specific case to make viable inferencesabout just what processes may have led to the products at hand. In the studiesof De Angelis and Selinker (2001) and Herwig (2001), there is scant mention ofthe notion of overgeneralisation; yet some of the examples discussed suggestprocesses of analogy involving both the target language and some languageother than the native in the acquisition of L3 or L4. For instance, Herwig notesa case of the use of the lexical hybrid foresld'gger by an English speaker appar-ently unaware of the precise form foresldr ('suggest') in the target language,Swedish. The transfer here comes from L2 German, but is not the simple sub-stitution of the German form (vorschlagen). It is beyond the scope of this paperto attempt to detail the exact processes that lead to forms such as foresld'gger, butclearly some kind of overgeneralisation in conjunction with transfer seemsmost plausible. If so, a good deal of work remains to be done on detailingthe range of possible interactions leading to such hybrids.

Our own study differs from those discussed in this section and elsewhereprimarily in terms of our focus on one source language, Swedish, with the dif-fering outcomes of cross-linguistic influence involving this source. Yet our find-ings concur in significant ways with those of investigations looking at languagedistance and proficiency as factors promoting transfer. Although the study ofthe interactions of such factors remains in its early stages, one generalisationseems possible. Whether researchers examine the effects of a single sourcelanguage (as we did) or compare the effects of two (or more) source languages,the outcome of cross-linguistic influence involves more than what a traditionalcontrastive analysis might predict. The interaction of factors such as proficiencywith structural contrasts can lead to a highly diverse set of outcomes.

AcknowledgementsWe would like to express our thanks to two anonymous referees for their

constructive criticism and also to Patsy Lightbown for her feedback on anearlier version of the paper.

CorrespondenceAny correspondence should be directed to Terence Odlin, Department of

English, Ohio State University, 421 Denney Hall, 164 West 17th Avenue,Columbus, OH 43210-1370, USA ([email protected]).

ReferencesAhukanna, J., Lund, N. and Gentile, J.R. (1981) Inter- and intra-lingual effects in learn-

ing a third language. Modern Language Journal 65, 281-287.Bjorklund, S. and Suni, I. (2000) The role of English as L3 in a Swedish immersion pro-

gram in Finland. In J. Cenoz and U. Jessner (eds) English in Europe: The Acquisition ofa Third Language (pp. 198-221). Clevedon: Multilingual Matters.

140 The international Journal of Multilingualism

Cenoz, J. (2001) The effect of linguistic distance, L2 status, and age on cross-linguisticinfluence in third language acquisition. In J. Cenoz, B. Hufeisen and U. Jessner(eds) Cross-linguistic Influence in Third Language Acquisition: Psycholinguistic Perspec-tives (pp. 8-20). Clevedon; Multilingual Matters.

Chumbow, B. (1981) The mother tongue hypothesis in a multilingual setting. In J-G.Savard and L. Laforge (eds) Proceedings of the Fifth Congress of the International Associ-ation of Applied Linguistics. Quebec: Laval tJniversity Press.

De Angelis, G. and Selinker, L. (2001) Interlanguage transfer and competing linguisticsystems in the multilingual mind. In J. Cenoz, B. Hufeisen and U. Jessner (eds)Cross-linguistic Influence in Third Language Acquisition: Psycholinguistic Perspectives(pp. 42-58). Clevedon: Multilingual Matters.

Dewaele, J-M. (1998) Lexical inventions: French interlanguage as L2 versus L3. AppliedLinguistics 19, 471-490.

Dewaele, J-M. (2001) Activation or inhibition? The interaction of Ll, L2 and L3 on thelanguage mode continuun:i. In J. Cenoz, B. Hufeisen and U. Jessner (eds)Cross-linguistic Influence in Third iMnguage Acquisition: Psycholinguistic Perspectives(pp. 69-89). Clevedon: Multilingual Matters.

Duskova, L. (1984) Similarity - an aid or hindrance in foreign language learning? FoliaLinguistica 18, 103-115.

Hammarberg, B. (2001) Roles of Ll and L2 in L3 production and acquisition. In J. Cenoz,B. Hufeisen and U. Jessner (eds) Cross-linguistic Influence in Third Language Acquisition:Psycholinguistic Perspectives (pp. 21-41). Clevedon: Multilingual Matters.

Herwig, A. (2001) Plurilingual lexical organisation: Evidence from lexical processing inL1-L2-L3-L4 translation. In J. Cenoz, B. Hufeisen and U. Jessner (eds) Cross-linguis-tic Influence in Third Language Acquisition: Psycholinguistic Perspectives (pp. 115-137).Clevedon: Multilingual Matters.

Jarvis, S. (1998) Conceptual Transfer in the Interlingual Lexicon. Bloomington, IN: IndianaUniversity Linguistics Club Publications.

Jarvis, S. and Odlin, T. (2000) Morphological type, spatial reference, and languagetransfer. Studies in Second Language Acquisition 22, 535-556.

Jordens, P. (1977) Rules, intuitions, and strategies of foreign language learning. Interlan-guage Studies Bulletin 2, 5-76.

Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. (1985) A Comprehensive Grammar ofthe English Language. London: Longman.

Ringbom, H. (1987) The Role of the First Language in Foreign Language Learning. Clevedon:Multilingual Matters.

Ringbom, H. (2001) Lexical transfer in L3 production. In J. Cenoz, B. Hufeisen andU. Jessner (eds) Cross-linguistic Influence in Third Language Acquisition: Psycholinguis-tic Perspectives (pp. 59-68). Clevedon: Multilingual Matters.

Schmidt, R. and Frota, S. (1986) Developing basic conversational ability in a secondlanguage: A case study of an adult learner. In R. Day (ed.) Talking to Learn: Conver-sation in Second Language Acquisition (pp. 237-326). Rowley, MA: Newbury House.

Selinker, L. and Baumgartner-Cohen, B. (1995) Multiple language acquisition: 'Damn it,why can't I keep these two languages apart?' Language, Culture, and Curriculum 8,115-121.

Singleton, D. (1987) Mother and other tongue influence on learner French. Studies inSecond Language Acquisition 9, 327-345.

Taylor, B. (1975) The use of overgeneralisation and transfer as learning strategies byelementary and intermediate students of ESL. Language Learning 25, 73-107.

Williams, S. and Hammarberg, B. (1998) Language switches in L3 production: Implica-tions for a polyglot speaking model. Applied Linguistics 19, 295-333.