Additive scope particles and anaphoric linkage in narrative and descriptive texts: a developmental study in French L1 & L2

Sandra Benazzo (UMR 8163 Savoirs, Textes, Langage & Université Lille 3), †Clive Perdue & Marzena Watorek (UMR Structures formelles du langage, CNRS & Université Paris 8)

Preliminary version of the article to be published in : M. Watorek, S. Benazzo & M. Hickmann (éds.) Comparative Perspectives to Language Acquisition: A tribute to Clive Perdue, Clevedon, Multilingual Matters.

1

Additive scope particles and anaphoric linkage in narrative and descriptive texts: a developmental study in French L1 & L2

Sandra Benazzo (UMR 8163 Savoirs, Textes, Langage, CNRS & Université Lille 3), †Clive Perdue & Marzena Watorek (UMR Structures formelles du langage, CNRS & Université Paris 8)

The core ideas of this article have been initally worked out for a joint presentation (Euresco conference « The Structure of learner language », Kolymbari, 7-12 October 2002), and afterwards refined in the form of the present paper, which was written before Clive passed away. We decided to publish it here1, as a tribute to Clive, in memory of a deep friendship and a fruitful collaboration, made of stimulating scientific discussions, hard work but also a lot of fun !

1. Introduction This paper addresses the general question of discourse construction from the specific point of view of the speaker/learner's use of additive words in the production of two different text-types – picture descriptions and retellings. These tasks will be described in detail in sections 2.1 and 3.1. We compare performances on these two tasks by different types of speaker: children aged 4, 7 and 10 years, adults speaking in their L1, and adult learners of an L2. The language spoken is French in all cases. We analyse these speakers’ use of ‘additive words’: additive scope particles aussi (also), encore (another), and temporal adverbs of contrast encore, toujours (again - still). Aussi associates with a constituent of the utterance it occurs in, and adds this constituent to an already established (set of) alternative(s). In what follows, we call this constituent the ‘scope domain’ of the particle. For encore, we understand the idea of repetition or addition, a meaning close to that of aussi. Encore differs however from aussi in that the set of alternatives are tokens of the same type (= another or one more). Encore also has two temporal values: (a) often associated with une fois – encore une fois (‘once again’) – it has a value of repetition: the same type of situation takes place at a later time; (b) it relates two temporal reference points: a state or action remains valid from a previous time to the time in question. This temporal value comes close to one of the values of toujours, whose meaning is usually described as marking (a) a permanent validity for the time span in question, or (b) the persistence of the validity of a state or action until a time in question. (b) is strongly reminiscent of encore, and we propose the following common denominator: “an action or state is valid for a time span t and for at least one other time span preceding t”. Used in (adult) discourse, all these words refer back, i.e. anaphorically, to a previous time span or a previous member of a set. The scope particles are attested early in the spontaneous production both of children (Gayraud 2004) and adult learners (Perdue, Benazzo & Giuliano 2002), whereas the temporal adverbs occur

2

somewhat later. What we attempt to do here is to analyse their role in organising the different text-types. The following table gives the sum of the tokens of these words found in the production of each group of speaker. Table 1: Tokens of words across groups and text types Particle

Aussi Encore toujours Total

Additive task children adults (15subj./gr) (10subj./gr.) 4 yrs 7 yrs 10 yrs L1 L2 13 12 9 9 18 26 15 12 3 17 2 15 15 39 2 41 42 36 51 41

tot. 61 73 73 207

4 yrs 19 18 0 37

Description children adults (20subj./gr) (10 subj./gr.) 7 yrs 10 yrs L1 L2 23 27 18 27 3 17 19 2 0 0 19 0 26 44 56 29

tot tot. 114 59 19 192

The table also shows that for the additive task, each group of children contained 15 subjects, and for the description, 20 subjects. 10 native adults performed both tasks. The group of adult L2 speakers (age range 20 to 46 years) was composed of 10 subjects whose L1 is Polish. All were recorded in France, after a length of stay ranging from 1 to 6 months (with the exception of one subject who has been in France for 2 years). Two of these learners, Waldemar and Ewa, were at a basic level of proficiency (Klein & Perdue, 1997), whilst the other eight were more advanced, using (unsystematically) some verbs in the present tense and with one or two past forms (passé composé) and the modal vouloir (‘want’); one also observes some cases of subordination with parce que (‘because’) et pour (‘in order to’).

2. Additive task: reference to people and time 2.1. The task For this task the subject was asked to produce an on-line narration based on a series of 30 pictures. The interviewer comments on the first four pictures, which serve as an orientation. The first introduces the setting: a small village with a church, a café and a general store, with in the background a castle atop a hill. The second allows the interviewer to give the point of the story, which is to help the king discover who has released the princess, who had been imprisoned in the castle. Only the subject can do this, as s/he has been watching what has been going on in the village all day. Pictures 3 and 4 introduce the possible candidates for the reward: Mr Red and Mr Blue (henceforth MrR and MrB). With the personal, temporal and spatial frame set up, the subject takes over and reconstructs from the following pictures, shown one by one, what MrR and MrB did on the day the princess was freed. The task is an adaptation of the “additive task” conceived first of all for adults2 by Dimroth (2002) specifically to elicit additive particles. Table 2 on the following page summarises the content of each picture and specifies the contexts where use of additive particles could be expected (“additive episodes” I-V). The additive episodes are organised in the following way:

3

175 132 92 399

Episodes I and V show a series of actions undertaken first by MrR and later by MrB. The same actions accomplished by different protagonists was designed to elicit aussi (MrB buys a ladder and MrR aussi buys a ladder), or some contextual equivalent such as également or comme MR (like MrR). Such expressions explicitly mark the additive relation between the protagonists for the same action-type. Episode II shows MrR at a café table who drinks one orange juice, then a second, then a third. The same protagonist performs the same action-type at successive time intervals. Here, the subject can either prolong the time interval: il boit encore/toujours, or focus on the different drink-tokens: il boit un jus... il boit encore un jus. Episodes III and IV are understood in relation to the church clock: at 3.30 MrR is sleeping, at 4.00 he is sleeping, at 4.30 he is reading a paper and at 5.00 he is sleeping. This episode requires both an additive relation in the domain of time as the same activity continues il dort encore/toujours, and the repetition of the same situation - he goes back to sleep: il dort encore/ de nouveau, or il re-dort. Marking these additive relations on protagonists or other entities, or on time intervals, contributes to discourse cohesion by building up anaphoric chains. We start by examining when these additive relations are indeed explicitly marked.

Table 2: The "additive-elicitation task" (adapted from Dimroth, 2002) Add. Episode

N° picture

Situation

I

01 02 03 04 05 06

introduction "village street" introduction "princess" introduction "Monsieur Rouge" introduction "Monsieur Bleu" R & B are in front of the church B leaves R remains in front of the church R leaves / walks towards the café R drinks an orange juice R drinks an orange juice R drinks an orange juice R leaves B walks along the street B smokes a cigarette B leaves in the bus R is seated on a bench 15.00 R sleeps on the bench 15.30 R sleeps on the bench 16.00 R sits up, reads the paper 16.30 R sleeps on the bench 17.00 R goes into a shop R comes out with a ladder R walks towards the hill B arrives back with the bus B walks towards the café B drinks an orange juice B goes into a shop B comes out with a ladder B walks towards the hill B walks up the hill towards the castle R stays near the trees B nears the castle R picks apples

II

III

IV

V

07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Expected additive item

4

aussi encore un / un autre encore un / un autre

encore / toujours encore / de nouveau

aussi aussi aussi aussi

2.2 Results Marking vs non-marking of additive relations For the L1 speakers, explicit marking increases with age. This increase is true of all episodes, but in different proportions: the three groups of children mark additive relations for entities less than those for time intervals. Even at 10 years, although almost 100% of the temporal relations are marked, only 4 subjects out of 15 (26.6%) mark the additive relation on entities. Indeed, the childrens’ results from episodes I and V makes one wonder whether it is at all necessary to mark these additive relations. The L1 adults do however overwhelmingly (80%) mark episode V, but not episode I (20%). Episode I is therefore not an obligatory context. This is so because a contrast other than that between the two protagonists is possible: both leave the scene, but MrB exits stage left whereas MrR exits stage right. The adults all mark the temporal relation of continuation, and 9 out of 10 mark the repetition. A preliminary comparison with the adult L2 learners shows that they behave more like the L1 adults than the children: only 20% explicitly mark episode I, but a large majority mark all the other additive contexts (60-90%, depending on the episode). The episode II has not been quantified, as too many different interpretations have been attested for this scene. Table 3 : Explicit marking of additive relations Additive Episode I V

III IV

Add. of Entities Add. of Entities

Add. of Time intervals (Continuation) Add. of Time intervals (Repetition)

4 year olds

7 year olds

10 year olds

Adults

L2

13,3 %

26,6 %

6,6 %

20 %

20 %

13,3 %

20 %

26,6 %

80 %

80 %

13,3 %

73,3 %

100 %

100 %

90 %

60 %

66,6 %

93,3 %

90 %

60 %

Linguistic means used In this section we examine group by group all the explicit means, whether they be direct or indirect, that subjects use to mark the additive relations. The four year olds The main characteristic of this group is the use of very simple utterances, by means of which children comment on each picture separately, rather than linking them to construct a story. Most children use deictic means to refer directly to the picture and the interviewer has to intervene to have the protagonists specified. In (1), the interviewer’s words are in bold:

5

(1)3

y a un monsieur qui va par là y a un monsieur qui va par là et après il vient encore et puis il boit son jus son truc c'est qui? Monsieur Rouge (Axelle)

there’s a man who goes that way there’s a man who goes that way and after, he comes again and then he drinks his juice his thing it’s who? Mr Red

et après il arrive au château and after, he arrives at the castle il coupe l'arbre lui ! he cuts the tree, him! Monsieur rouge? MrR? oui Monsieur Rouge yes MrR Monsieur Bleu il va couper cet arbre là (Car.) MrB he goes to cut that tree there •

Episodes I & V: protagonists

The explicit marking of the additive relation is rare overall, but when it is marked, all subjects use aussi. But there are problems with its interpretation. Either the referent in question is undecidable, as in (1), in which case the scope domain of aussi is an uninterpretable pronoun, or the position of aussi is compatible with two scope domains. When aussi is in post-verbal or utterance-final position, its scope domain can be either the subject or the predicate of the utterance: (2)

après il ... va acheter une échelle aussi after, he goes to buy a ladder too

(Mélanie)

In (2), one needs the preceding picture to be able to decide whether aussi’s scope domain is "il" (addition of protagonists) or "acheter une échelle" (addition of actions). We also find the particle at the beginning of the utterance: (3)4

M.Bleu boiva aussi un verre... et aussi M.Bleu prena l’échelle(Lena)

MrB drinked also a glass and also MrB taked the ladder

(3) describes the episode (V) where MrB is recapitulating MrR’s actions: the informational context thus indicates that the scope domain of aussi in bold type is MrB (underlined in the gloss). This position for this scope domain is in fact impossible in TL-French. The vast majority of these interpretation difficulties can be attributed to the fact that these children do not take into account the state of knowledge of the interviewer.

•

Episodes III & IV: situations

The continuation of the action of sleeping in episode III is often marked by temporal adverbs: encore (‘still’, 6/15) or longtemps (‘for a long time’, 1/15). But ‘indirect means’, such as aussi or the negation, are also often present. (4)

là il est allongé / là aussi (Lucie)

there, he is stretched out / there too

Lucie uses aussi to express a similarity between two pictures, to which she refers deictically. Examples such as (4) clearly show that at this age, children can produce descriptions when performing this narrative

6

task. In (5), Pauline, in commenting successive pictures, indirectly asserts that the action is continuing by denying that it has stopped: (5)

il boit / il a pas fini (de boire) (Pauline)

he drinks/ he has not finished (drinking)

A last noteworthy fact in this group’s expression of continuation is that the concept is sometimes confused with that of iteration, insofar as encore is used with a verb expressing a telic event to comment a picture depicting a continuative rather than a repeated situation. (6)

il s’allonge / il s’allonge encore(Lola.)

he lies down/ he lies down still

Another possible interpretation is that, by analogy with là aussi used as a deictic commentary, Lola is commenting the repetition of a scene she has already observed: encore une fois je vois quelqu' un qui s'allonge ‘again I see someone lying down’. Otherwise, the expression of repetition is unproblematic, with 6 subjects out of 15 using the specific prefix re- as in this example from Quentin: (7)

il dort

/ il redort (Quentin)

he goes to sleep/ he re-goes to sleep

The seven year olds With the exception of là (‘there’, referring to the picture) all the deictic uses of the 4-yrs production have disappeared. The protagonists are clearly identified, with in some cases an expression which is overexplicit in relation to adult productions (see example 8: the MR he ... the MB he.. , and with other children: MR he... or there is MR who..., in a context of reference maintenance). (8)

le M.Rouge il rentre dans le magasin d'outils le M.Rouge il ressort avec une échelle (Jus.)

the MrR he goes into the tool shop the MrR he comes back out with a ladder

The strong pronoun lui is used as in TL French to contrast two topics (5 subjects out of 15): (9)

M.Rouge va prendre des pommes et M.Bleu lui suit le chemin (Bas.)

MrR goes to take some apples MrB (as for) him continues along the path

Use of re- is now extended to denote the restitutive variant of repetition (ressortir,’come back out of’, see example 8, revenir, ‘return’, etc.).

•

Episodes I & V: protagonists

Aussi remains the preferred linguistic means for marking this additive relation, and now the scope domain of the particle has become clearly identifiable thanks to the use of the contrastive pronoun lui. The means are now in place for (over-)marking contrastive topics, as (10) shows: (10)

il est parti pour aller à l’épicerie...

he’s gone to the grocer’s

7

et lui aussi il a pris une échelle (Mél.)

and he too he has taken a ladder

là M.Bleu il part et M.Rouge aussi il part (Mat.)

there MrB he leaves and MrR too he leaves

Other means can now be used to mark this relation, as with the explicit comparison of (11): (11)

•

il va vers le chemin comme son frère (Dap.) he goes toward the path like his brother

Episodes III & IV: situations

Indirect means such as negation or the described similarity between two pictures (examples 4 and 5) are still used to mark continuity, as is encore (5 subjects). However, other means are also used: toujours (3 subjects), and the verb continuer (1 subject): (12)

et puis il est toujours en train de dormir (Cor.) and then he’s still sleeping après il continue à dormir (Mat.) after, he continues to sleep

although the use of this verb in an iterative context (13) suggests that some 7-yrs still have problems in differentiating continuity and iteration: (13)

il lit un journal / il continue de dormir comme tout à l’heure* he reads a paper / he continues to sleep like shortly before

(Bla.)

Other subjects use however the re-V prefix in this iterative context.

The ten/eleven year olds More sophisticated means are used to contrast the two protagonists, such as a less rigid and TL-like use of contrastive topic, which can be associated with the successful contrast of (simultaneous) actions by means of tandis que (‘whereas’) and pendant que (‘while’): (14)

M.Bleu lui monte vers le château... MrB HE goes to the castle tandis que M.Rouge a déjà cueilli des pommes (Luc.) whereas MrR has already picked apples et M.Bleu il s' en va en voiture tandis que M.Rouge # il est assis sur un banc ... M.Rouge appuie son échelle contre un pommier où y a des belles pommes rouges pendant que M. Bleu arrive (Axe.)

and MrB he goes away in a car whereas MrR # he is seated on a bench MrR leans his ladder against an apple tree where there are nice red apples while MrB arrives

il rencontre M.Rouge qui cueille des pommes lui il monte encore plus vers le château(Eme)

he meets MrR who is picking apples HE he climbs still further towards the castle

pendant ce temps M.Bleu arrive (Nic.)

during which time MrB arrives

8

•

Episodes I & V: protagonists

The contrastive pronoun lui + aussi structure used systematically by the 7-yrs in initial position can also be used as an antitopic by the 10-yrs. For topic continuity, the 10-yrs, unlike the younger children, also use ellipsis of the subject (zero anaphor), as the following example shows: (15)

après il va voir vers le magasin d’outils et en sort lui aussi avec une échelle(Eme)

after, he goes to see in the tool shop and ∅ comes back out he too with a ladder

The means marking addition of entities become richer, as does the organisation of information. The 10yrs use a procedure also used by adults which consists in framing a series of actions by stating they are similar – comme - to a preceding series, then giving the detail. Here is an adult version of this procedure: (16)

et là il fait comme M.Rouge tout à l'heure il s'assoit à la terrassse du café et il boit un coup ensuite il rentre dans le magasin (Tif.)

and there he does like MrR did before he sits down at the café terrasse and he has a drink then he goes into the shop

The comparison with native adults shows that these have an even bigger repertoire for expressing the additive relation, using également and à son tour as well, as in the following examples: (17)

il en ressort avec une échelle sur le dos he comes back out with a ladder on his back et ∅ se dirige lui aussi à son tour vers le château (Bru.) and ∅ directs his steps he too in turn towards the castle il boit également une orangeade he equally drinks an orangeade il va également acheter sa petite échelle (Lin.) he equally goes to buy his little ladder

Episodes III & IV: situations To express continuity, the 10-yrs use a variety of means: encore (6/15), toujours (6/15), continuer (1/15); the prefix re-V (il se rendort, ‘he goes back to sleep’, and, once il recommence à dormir ‘he starts sleeping again’) is the only means now used to mark reiteration. The Polish learners of L2 French As we said, the explicit marking of additive relations is quantitatively like that of the native adults; these adult learners systematically attempt to give structure to their reference to the protagonists, and to the temporal relations. Recall also that this group does not master the verb morphology of French, hence, some of the English glosses given below contain verbs in the infinitive form, so as not to over-interpret the learners’ underlying knowledge.

9

•

Episodes I & V: protagonists

(18) gives the complete repertoire of these learners’ means for marking additive relations: aussi, comme, même and the idiosyncratic à nouveau (used by Agnieska to convey ‘also’ rather than ‘again’) (18)

M.Bleu parti / M.Rouge aussi partir (Ewa) MrB leave/ MrR also leave M.Bleu il assis comme avant M.Rouge MrB he seated like before MrR et il boire quelquechose comme avant M.Rouge (Barbara) and he drink something like before MrR il prend le même passage petit à château (Pawel) he take the same passage little to castle M.Bleu à nouveau* boit (Agnes) MrB again* (=also) drink

The beginning learners tend to place the particle aussi immediately to the right of its scope domain (as Ewa, in 19), whereas the more advanced learners use the post-verbal/final position (as Agn., in 19). (19)

M.Rouge juste à côté M.Bleu il aussi devant l'église M.Bleu parti / M.Rouge aussi partir (Ewa)

MrR just next to MrB he too in front of the church MrB left/MrR also leave

après M.Rouge aller parti aussi (Agnes)

after, MrR go leave too

Almost as if they were aware of the different scope possibilities of aussi in this position, these learners add supplementary means (comme M.Rouge, or, as in the following example, même ‘the same’) in order to make explicit the scope domain of aussi: (20)

après il va au magasin. il il va acheter aussi l’échelle euh aussi il va acheter achète échelle il va aussi le même chemin M.Rouge pour le château(Edyta)

after, he goes to the shop he he goes to buy also the ladder er also he goes to buy buys ladder he goes also the same path MrR for the castle

What is remarkably different with the adult learners is that none uses either the French-specific contrastive pronoun lui, or, therefore, the contrastive topic structure lui aussi.

•

Episodes

III & IV: situations

The means used by the beginners tend to confirm the results of other studies (Benazzo, 2000; 2003) that additive aussi is used/acquired before the temporal adverbs of contrast. As we have seen, Ewa has aussi, but does not use encore. The beginners resort, completely unlike the children, to a chronology (21) with or without the aid of the church clock, or, somewhat like the young children, to the expression of similarity between pictures (22): (21)

il dort il dort à 30 minutes à 5 h il fatigué et il dort (Ewa)

he sleeps he sleeps to 30 minutes at 5 o’clock he tired and he sleeps

(22)

M.Rouge se couche devant l'église ici il se couche aussi (Mat.)

MrR lie down in front of the church here he lie down too

10

The intermediate learners do mark the addition of time intervals, using encore both for continuity and for iteration. (28)

il dort sur le banc / il dort encore (Pawel)

he sleeps on the bench/ he sleeps still

The positioning of this adverb, and also of toujours, is however unstable: 5 learners place them pre- and post-verbally: (29)

MR dort à côté église / encore dort MR lit le journal ou livre il dort encore (Renata)

MrR sleeps next to church/still sleeps MrR reads the paper or book he sleeps again

il est toujours à côté de l'église... il toujours dort (Barbara)

he is still next to the church he still sleeps

il dort encore M.Rouge il a [revej].... ensuite il M.Rouge encore il dort (Edyta)

he sleeps still MrR he has awake then he MrR again he sleeps

It may be that the position of encore varies with the continuity/iteration distinction, and that the adverbs behave differently with respect to the copula and main verbs, but the data are too scant to be affirmative. What can be reliably asserted, though, is that none of the adult learners uses the French-specific prefix reV. Comparison L1-L2 We can summarise the development observed in the child learners in the following way: -

Protagonists

The 4-year-olds mainly use aussi in association with extra-linguistic referents (là aussi, ‘there too’) before using it endophorically as a means to reinforce text cohesion, i.e., as an anaphor. Its scope domain becomes clearly identifiable from 7-yrs onwards, thanks to the emergence of the contrastive use of the strong pronoun lui. These two words together (lui aussi) function to link actions by contrasting their actants. This development parallels that in reference to people: NP-expressions are ambiguous in the 4-yrolds’ production, as they do not take account of the interlocutor’s state of knowledge; forms are on the contrary overexplicit at age 7, and contextually adjusted at age 10, with the emergence of zero anaphor. -

Time intervals

In this domain, we see the progressive emergence of specific and adequate means for marking the additive relation. 4-yr-olds mark continuity indirectly (negation, aussi accompanied with a deictic expression) before the subsequent use of encore, toujours, and the verb continuer. The iterative value comes systematically to be marked by the specific prefix re-. The two values continuity and iteration are not systematically differentiated even with some 7-yr-olds, as the incompatibilities between encore, continuer and the Aktionsart of the associated predicate attest.

11

-

Form-function relations

All the elicited additive items do seem to be part of the youngest childrens’ repertoire, and what has to be acquired is their discourse function. The same reasoning holds also for lui: the strong pronoun is used, albeit rarely, in the youngest group, but with either a deictic or a non-contrastive value (respectively il s’en va, lui in example 1, and il a un verre devant lui ‘he has a glass in front of him’). The contrastive function is not attested before age 7. For the adult learners we observe systematic attempts to mark additive relations, differently from the children. What differentiates the adult learners from the adult L1 speakers is the linguistic means employed. In particular the specific means to mark contrastive topics lui (aussi) or repetition/restitution – re- - are not acquired; the additive particles aussi/encore, or indirect means are used instead, and in this particular respect we do see similarities with the productions of the very young children.

3. The descriptive task: reference to entities and places 3.1 The task The descriptive task involves describing as completely and precisely as possible the spatial structure of a poster depicting an urban scene, and the positions of entities relative to each other (what is where in relation to what). The children were asked to describe the poster to a person who was present but could not see it, whereas the adults were told that their recording would be used in a further part of the experiment5, to see if another person could reproduce the poster from the recorded description, by drawing it. The procedure thus gives more freedom to the subject in contrast to the additive task, as s/he can choose the point of departure and the linearisation strategy. The poster did however provide a constant set of spatial configurations to be described. These configurations are described in what follows as a relation between a relatum (or ‘ground’, in Talmy’s, 1985, terms) and a theme (‘figure’, in Talmy’s terms). As we did in 1., we first take the childrens’ productions, then compare them with those of the adult learners.

3.2 The results The four year olds The young children found this task extremely difficult, and we find few theme-relatum relations marked explicitly in their productions. The general strategy is to take the poster as an implicit relatum, and then provide a list of themes. When the relatum is expressed, the order theme + relatum is systematically observed. A majority of utterances thus consist of a bare NP, to which an additive particle

12

may be associated left- or right-adjacent. The particle also associates with a deictic spatial adverb, allowing the child to express that another token of the same entity-type can be seen in another place. - Addition of entities is marked by aussi and encore Quentin’s production illustrates the minimal strategy employed by this age group: the poster serves as implicit relatum, and the different themes are simply enumerated, with aussi marking the end of the list: (30)

une voiture un trottoir et aussi des oiseaux (Que.)

a car a pavement and also some birds

Even when the relatum is made explicit, aussi can still be interpreted as having the same function, as its exact scope domain is otherwise hard to identify: (31)

un vélo vers l’arbre ya une petite maison un camion sur la route aussi (Alex)

a bicycle towards the tree there’s a small house a lorry on the road also

Encore creates an additive link between tokens of the same type of theme located in respect to different (implicitly or explicitly expressed) relata. Antoine uses the deictic là to denote the two relata: (32)

et après une coccinelle là et encore une coccinelle là(Antoine)

and after, a beetle there and another beetle there

Encore is never used with a temporal value at this age, but may take a meta-discursive value: ‘moreover’.

- Addition of places is marked by aussi This additive link is used exclusively with the deictic adverbs of place ici and là: (33)

et puis il y a trois fenêtres ici aussi et là aussi (Man.)

and then there are three windows here too and there too

Encore is however never used at this age to mark continuity (extension) of the relatum. The seven year olds As we saw in the additive task, the 7-yr-olds produce very regular discourse, with information in both domains of entities and space being clearly organised. Utterances now more systematically contain reference to the theme and the relatum, in either order. If the relatum is left implicit, this is now not only in relation to the poster in its entirety, but also to salient sub-parts of the poster. If the relatum is (the space occupied by) an object, then reference maintenance is explicitly (and redundantly) marked (ex. 34), giving an effect of "over-regularisation" which disappears in the 10-yr-old production, as we shall see.

13

(34)

il y a un tabac au milieu de l’avenue... there’s a tobacconist’s in the middle of the avenue il y a une statue aussi au milieu de l’avenue (Corinne) there’s a statue as well in the middle of the avenue

Utterances most often contain an existential verb, and aussi is placed after this verb when used. - Addition of entities is marked by aussi and encore, and addition of places by aussi. There are still traces of the 4-yr-olds’ ‘listing’ strategy with aussi being used to close off the list: (35)

il y a plein de gens il y a un bureau de tabac il y a des papeteries aussi des hôtels (Rém.)

there are lots of people there’s a tobacconist’s there are stationer’s shops also hotels

and encore is still used to create an additive link between tokens of the same type of theme, rather than to express continuity. Corentin’s example above (34) shows however another entity located in relation to an explicitly maintained relatum, and the following example of Alexandra shows a mirror organisation, with aussi creating a link between different relata associated with the same theme-type, the explicitly maintained blocks of flats: (36)

à l’angle de l’affiche il y a des immeubles at the corner of the poster there are blocks of flats et de l’autre côté de l’affiche il y a aussi des immeubles (Ale.) and on the other side of the poster there are also flats

The ten year olds The main difference between this group and the 7-yr-olds is that now, information from the spatial domain is more highly structured, with the spatial relation between relata being expressed. Utterances are now mainly of the form: PP – Vexist – NP, expressing the informational order Relatum + Theme.

- Addition of entities is marked by aussi and encore If the relatum is inferrable from the co-text, it is now left implicit rather than being explicitly maintained, as with the 7-yr-olds, cf. example (34): (37)

en bas de cet immeuble il y a une rue il y a aussi une fontaine (Ana.)

at the foot of the block of flats there’s a street there’s also a fountain

- Addition of places is marked by aussi When two relata are associated with the same clearly inferrable theme type, then this theme-type may be left implicit (as a mirror image of 37): (38)

à droite il y a des arbres derrière aussi (Bas.)

to the right there are trees behind as well

14

- Continuity of space is marked by encore This relation is only attested with the 10-yr-olds and the adults. The children use encore: (39)

encore en dessous à droite de la fontaine il y a un banc further below to the right of the fountain there’s a bench encore plus bas il y a un monsieur (Léa) still further down there’s a gentleman

whereas the adults use both encore and the (temporal) toujours (8 sujets sur 10) : (40)

ensuite toujours dans la profondeur on a un autre immeuble (Linda) then still deep in the background we have another block of flats

The Polish learners of L2 French Ewa and Waldemar, the beginners, have a poor repertoire of spatial expressions, often use nouns without articles, and use the idiosyncratic c’est, est as an existential. The only particle attested – aussi – is placed post-verbally. In the following example, it links two relata in respect to a similar, explicitly-maintained theme-type: (41)

devant cette camion c’est petit la voiture derrière c’est aussi petit la voiture (Ewa)

in front of that lorry it’s little the car behind it’s also little the car

Reference maintenance is however generally left implicit, as is the case with the relatum entities in (42), which illustrates a relatively well-organised and comprehensible discourse using minimal means: (42)

à côté de la rue à gauche il y a euh maison et euh à côté il y a tabac et cinq arbres et la fontaine à droite il y a deux voitures et une femme avec bicycle et derrière il y a deux maisons (Klaudia)

next to the street to the left there is er house and next to Ø there is tobacconist and five trees and the fountain to the right (of Ø) there are two cars and a woman with bicycle and behind Ø there are two houses

The intermediate learners tend on the contrary to (over-)mark reference maintenance explicitly, much in the same way as we saw with the 7-yr-olds (cf. 34): (43)

dans la place il y a beaucoup de personnes dans la place il y a aussi un vélo (Ren.)

in the square there are lots of people in the square there is also a bicycle

The intermediate learners tend to use il y a rather than c’est, and their utterances regularly show the form: PP – il y a – NP, with aussi placed post-verbally when used. Aussi is in fact used relatively frequently (26x) to mark the addition both of entities and of places: (44)

je vois aussi l’homme qui lit le journal à côté la rue je vois aussi un homme (Bar.)

I see also the man who reads the paper next the street I see also a man

15

Encore is used however neither to express addition of an entity-token, nor to express spatial continuity. The two occurrences of encore found in this sub-group have a meta-discursive function.

4. Summary and discussion We are making no new claims in this paper, but rather looking at some old problems from another point of view. We have found, as others have, that whilst there are some superficial similarities between child and adult learner production at points along the acquisition process, these similarities plausibly have different underlying causes, and are in any case outweighed by the differences. The additive means we have examined - aussi (également), encore, toujours - refer back to pieces of information already developed during the construction of discourse, linking them to pieces of information in the utterances in which they occur. In the texts examined here, this information comes from the referential domains of people, place and time (here, temporal reference and aktionsart). We have seen that the nature of the links established, changes over time for the child learners, whereas this not the case for the adults.

Summary - in the 4-year-olds' production, the problem for the analyst is to work out either (a) which is subject/agent of the actions mentioned – the relevant protagonists of the additive task are denoted with proper names or deictically used pronouns which are ambiguous half of the time (cf. also Hickmann 1982) - or (b) where the themes introduced in the description task are located - entities are introduced, appropriately, with indefinite articles but reference to these entities is never maintained and, in the domain of space, the poster serves as an implicit relatum for the themes, which are enumerated, simply evoked, or related to an almost unstructured spatial context (ici, là, en haut, en bas): a list of themes may be closed by aussi, two tokens of the same type of theme may be related by encore un or the spatial context may be related deictically là aussi. Thus almost the only trace of attempts to structure these configurations are the additive particles, which create links between the extra-linguistic theme or spatial context, but not the linguistic context. Moreover (c) the time-line in the story is non-existent, as the children seem to comment on each picture one by one. In these circumstances, the particles have a relatively considerable structuring power, although the deictic anchoring of these children’s production leads to an exophoric use of the particles. Aussi functions to link a similar referent-type (theme or process) to two different locations - là aussi (4, 33) or a similar action to two different protagonists - il s'en va aussi (2) -; encore serves, inappropriately, to link a type of process to a previously mentioned process-type - il s'allonge encore6 (6). (It seems that continuity and repetition are not well distinguished at this age).

16

- in the 7-year-olds' production, reference maintenance (to entities, space and time intervals) is a problem that is being worked on. Protagonists as well as locations are carefully distinguished: as in many previous studies, we have seen that the anaphoric expressions are even over-explicit in relation to adult native production. Aussi links a similar action to two different protagonists, sometimes redundantly (cf. ex. 10 : lui aussi il a pris une échelle....comme son frère). In the descriptive texts, the theme is localised more often, and this relatum is also specified redundantly in relation to adult production, when the simple additive particle localising two themes in relation to this relatum would have sufficed, as with the repetition of the relatum (cf. ex.34: il y a une statue aussi au milieu de l'avenue). The time-line in the story now emerges clearly, with deviations from the chronology beginning to be marked (avant de ‘before’, and one instance of pendant que – ‘during’). Repetition is consistently marked by re- (sometimes overmarked: il redort encore une fois), and continuation either by encore or by toujours. - the 10 -year-olds' production is fluent. The explicitness of the expressions referring to the protagonists is appropriate, which means that the functions of the NP + pro construction are now restricted to the reintroduction or contrast of one of the two protagonists, while mere subject maintenance can be appropriately expressed by zero anaphora (cf. ex. 15 : et ∅ en sort lui aussi avec une échelle). The domains of time and space become malleable as other conceptualisations than discreteness become available: means to express temporal simultaneity and temporal and spatial continuity are used unproblematically. Although the 4-year-olds and the adults produce very elementary utterances, the differences in organisation behind them is striking. In adult learner production, reference to protagonists relies on minimal means but is functional: there are no cases of unclear reference. Zero anaphor is used effectively right from the beginning, in alternation with pronoun, and contrasts with use of a full NP (proper name, or learner-specific je vois NP). Aussi is used in this context (cf. ex. 19: MR juste à côté MB / il aussi devant l'église). The strong pronoun lui is not attested, but there is attested some idiosyncratic use of NP + pro. There is a remarkable difference in the domains of space and time, which are structured right from the beginning, with the one exception of the absence of toujours/encore to express spatial continuity. Both temporal continuity and repetition are however expressed, even if most subjects use the multifunctional encore for both temporal contexts. The French-specific means for expressing repetition re-V - is not acquired. The most basic learner, Waldemar, has recourse to the deictic ici, but in stark contrast to the children, uses it as a context from which to build up a story: (45)

voilà l'église et trois heures et demie ici 4h MR dormi. (Wal.)

here’s the church and half past three here 4 o’clock MrR sleep

17

Table 4 attempts to visualise for the picture description task the cohesive capabilities of the children in respect to additive relations: Table 4: Children’s additive relations Description task

Additive task

Ref. to space

Ref. to entities

Ref. to time spans

Deictic anchoring : et puis là il y a 3 fenêtres ici aussi et là aussi

Deictic anchoring : là il est allongé // là aussi

Deictic anchoring : là il lit une histoire là il redort (dort encore)

Relatum ???: une voiture un trottoir et aussi des oiseaux

Protagonists ?? le monsieur (=MB) il s’en va il (=MR) s’en va aussi

Repetition of a situation or metadiscursive use of ‘encore’ ??

7 year olds

Explicit relatum c’est une ville où il y a plein de gens il y a un bureau de tabac il y a des papeteries aussi des hotels

Unambiguous maintenance and contrast : et lui aussi il a pris une échelle et il est allé lui aussi au château

Repetition vs. Continuity (same situation for consecutive time intervals) après il s’endort et puis il est toujours/encore en train de dormir

10 year olds

Space continuity en dessous à droite de la fontaine il y a un banc encore plus bas il y a un monsieur

4 year olds

zero anaphor: à droite il y a des arbres derrière aussi

Simultaneity (2 situations for the same time interval) MB se dirige vers le château pendant que MR est en train de cueillir des pommes zero anaphor: il va voir vers le magasin d’outils et ∅ en sort lui aussi avec une échelle

Discussion In their analysis of the early manifestations of finiteness-marking in child and adult learners, Dimroth et al. (2003) identify a common developmental stage – the so-called “conceptual ordering stage” - where the three information components of a finite declarative utterance are unambiguously realised. They term these components the topic, the predicate and the link. Under topic is included the expression(s) that refer(s) to the temporal and/or spatial and/or personal characteristics of the event that the utterance is about. The predicate7 denotes the state of affairs that is claimed to hold for the topic. The information unit called topic thus provides a (spatio-)temporal anchoring point for the information referred to in the predicate. The term anchoring is used here to cover the basic functions of topics, namely (a) the identification of what is talked about and (b) the embedding of the actual utterance in a discourse world. The discourse world can be text-internal (text, anaphor), or text-external (situation, deixis). Successful communication is dependent on the speaker's and hearer's mutual agreement on what is topical and what

18

is predicative information. Linking elements express a more abstract grammatical operation. Linking devices do not carry content information, their function is rather to express that the content of the predicate is not just a mere idea but claimed to hold for the topic situation in question. By executing a linking operation the speaker asserts that the predicate holds for the topic. All the utterances we have examined are to be interpreted as semantically finite: they assert the validity of a state of affairs with respect to a topic. Dimroth et al. (ibid.) found different ways to express linking during the acquisition process: it might be explicitly expressed by target language adequate (grammatical) means, or by intermediate (lexical) learner solutions. Among these intermediate solutions they identify feature precisely the additive particles we have been studying here. In a subsequent study, Dimroth (2002) demonstrates that the scope domain of these additive particles in the type of text we have been studying is typically (part of) the topic. In her own “additive elicitation task”, which we have adapted here, topic information typically concerns the protagonists and the temporal structure; in a description, the topic information typically concerns the relatum. Dimroth (2002) further relates one of the functions of the additive particles we have been studying to the linking operation: the (topic-related) particles express that a state of affairs under discussion applies to (part of) the topic component of an utterance. Both L1 and L2 data show in fact that there is a stage of development just in advance of the conceptual ordering stage when morphological finiteness marking, the TL-carrier of assertion, and additive particles are in complementary distribution. Dimroth (ibid.) discusses the following type of minimal pair: (46)

die sind runnergefallen von dem wagen und der mann auch runtagefallen

they have fallen out of the car and the man also fallen out

noting the correct morphological form of the finite verb – sind - in the first utterance, and its replacement by topic-related auch in the second utterance, and concluding that even in a relatively advanced learner variety such as this, “assertion does apparently not deserve an independent expression in utterances with … auch”. Nederstigt (2002) explains the same distribution (auch vs. finite verb form) in her child data in terms of a conflict: […] Caroline’s particle use clearly reflects a conflict between AUCH and the finite verb. This conflict between the location of the particle and the finite verb provides evidence for the analysis of AUCH as an overt assertion marker because without this function the emergence of the finite verb in AUCH-utterances would be without consequences. (Nederstigt 2002: 275) From this perspective, the simple assertion: ‘the state of affairs is valid for the topic (time, entity)’, has variants: it is valid for another topic entity (‘also’), or for another topic time (‘again’), etc. Dimroth calls this the “syntagmatic function” of the additive particles. This function is mastered by all the learners we have been studying: Nederstigt (2002) finds frequent examples as early as age 1;5 in German. But the nature of the topic information is very different between young and older children, and adults, and this what we have to explain.

19

The other function – Dimroth’s “paradigmatic function” – of the additive particles is to relate, in fact to contrast, an information unit of the present utterance with the same information unit of a previous utterance. In (46) the two topic entities die and der Mann are related: the man is explicitly added to the set of people who have fallen out of the car. It is in this paradigmatic dimension that we can place the development observed in this paper. Such development is well-researched in children for the domain of entities (Hickmann 1982, Hendriks 2000), but a parallel, less well-documented development is also valid for space and time. The observations made in this paper incite us to characterise the children’s developmental task as one of topic management. At 4 years, the topic component of children’s utterances is either left implicit or, for space and protagonists, expressed deictically. All stimuli are commented in coincidence with the time of utterance: the connector puis (‘then’), and sometimes encore, is used metadiscursively rather than to establish a chronology; the repetition of a state of affairs is contextualised by means of deictic spatial là (‘there’). At 7 years, we observe a regular structuring of the topic component in the domains of space, protagonists, and also time. In the descriptive texts, the topic space is divided into discrete sub-spaces, giving a preponderant topological relation of inclusion of the theme in the relatum. The protagonists in the retelling are carefully identified and contrasted by means of clitic and strong pronouns used anaphorically. In the same task, a chronology is established; puis is used anaphorically to link perfectively presented states of affairs. Then, the discrete relations between contexts can be supplemented by continuity: first, indirectly, in the domain of time, by use of adverbs that extend the time of situation to be valid for a further time span (encore, toujours), and only later, at 10 years, in the domain of space, where the relatum is extended, thanks to the use of these same adverbials: encore plus à droite, etc. These being anaphoric relations, it becomes clear why such temporal values occur relatively late in the acquisition process (cf. Introduction). So what we see in the children is a development from linking of extra-linguistic referents, to linking a theme or an event to a topic within an utterance, to crossutterance linking, as they come progressively to understand how to relate the information expressed in the topic of different utterances along the text. We have seen superficially similar ‘over-markings’ in the 7-yr-old and adult production, corresponding partially to what Ch. v Stutterheim and M. Watorek have called "prototypical processing" of these verbal tasks (cf. Watorek 1996). Both types of learners show this stage, but children have to get there. Adults know from their first language how to relate the information expressed in the topic of different utterances along the text. They acquire sufficient (if idiosyncratic) means to do so in another language, and if this functions - they may stop (cf. the absence of more specific morphosyntactic means as re-V). But we notice that not all the functions used by the 10 years old children are expressed by the adult learners studies here, which may provide a communicative motivation to acquire more.

20

References Becker, A. and Dietrich, R. (1996) The acquisition of scope in L2 German. LiLi 104, 115-40. Benazzo, S. (2002) Communicative potential vs. structural constraints: explanatory factors for the acquisition of scope particles. In S. Foster-Cohen, T. Ruthenberg & M.L. Poschen (ed.) Eurosla Yearbook 2 (pp. 187-204). Amsterdam: Benjamins. Benazzo S. (2003) The interaction between verb morphology and temporal adverbs of contrast. A longitudinal study in French, English and German L2 (pp. 187-210). In C. Dimroth and M. Starren (eds.) Information structure, linguistic structure and the dynamics of language acquisition. Amsterdam: Benjamins. Carroll, M. et al. (2000) The relevance of information organisation to second language studies. Studies in Second Language Acquisition 22, 441-66. Dimroth, Ch. (2002) Topics, assertions, and additive words: how L2 learners get from information structure to target-language syntax. Linguistics 40(4), 891-923. Gayraud, F. (2004) Emergence et développement du placement des particules de portée, Acquisition et Interaction en Langue Etrangère 20, 173-196. Hendriks, H. (2000) The acquisition of topic marking in L1 Chinese and L1 and L2 French. Studies in Second Language Acquisition 22, 369-397. Hickmann, M. (1982) The development of narrative skills. Pragmatic and metapragmatic aspects of discourse cohesion. Unpublished doctoral dissertation, University of Chicago. Jordens, P. (2002) Finiteness in early Dutch. Linguistics 40(4), 687-766. Klein, W. (1994) Time in Language. London: Routledge. Klein, W. and Perdue, C. (1997) The basic variety. Second Language Research 13(4), 301-48. Lasser, I. (1997) Finiteness in adult and child German. MPI Series in Psycholinguistics. Nijmegen: MaxPlanck-Institut. Nederstigt, U. (2002) The development of the focus particles AUCH and NOCH in German child language. PhD Thesis. Universitaet Humbolt, Berlin. Nederstigt, U. (2006) Additive particles and scope marking in child German. In G. van Geenhoven (ed.) Semantics in acquisition (pp.303-328). Dordrecht: Kluwer. Penner, Z., Tracy, R. and Wymann, K. (1999) Die Rolle des Focuspartikel auch im frühen kindlichen Lexikon. In J. Meibauer and M. Rottweiler (eds.) Das Lexikon im Spracherwerb (229-251). Tübingen and Basel: UTB-Frank. Perdue, C., Benazzo, S. and Giuliano, P. (2002) When finiteness gets marked. Linguistics 40(4), 849-90. Talmy, L. (1985) Lexicalization patterns: semantic structure in lexical forms. In T. Shopen et al. (eds.) Language Typology ans Syntactic Fieldwork, Vol. 3 (pp.57-149). Cambridge: Cambridge University Press. Watorek, M. (1996) Le traitement prototypique: définition et implications. Toegepaste Taalwetenschap in Artikelen 55, 187-200.

21

1

We thank Stephanie Haberzettl for making possible its publication in this volume.

2

In the first picture of the original story we see the princess with an axe in her skull, and the task is to find the

murderer.... 3

Learners are identified by their first names. Transcription conventions are kept to a minimum. [ ] enclose sequences

in broad phonetic transcription, and ... indicates the omission of an irrelevant passage. A self-interruption is marked by /, and a silent pause by #. In cases where the form used by a learner has a different semantic value from the targetlanguage (TL) value, it is followed by *. 4

The verbal forms used by Lena are “creative” (overgeneralisation of fr. passé simple) in that they do not correspond

to the TL, although such errors are very common at this age. We also note that in this example, we cannot exclude that the scope domain of aussi is the whole utterance: (metadiscursive) aussi = ‘moreover’. 5

This “further part” is in fact fictitious.

6

But repetition is mainly marked by re-.

7

Note that the term predicate is not used in its syntactic/categoric sense.

22