Reflective task 1
You probably guessed from the context that the German speaker was asking for something: a complete stranger is unlikely to invite you to a party or tell you the latest football results unprompted! He was in fact asking you to pass him the salt.

If you had actually been present in the situation, it is quite likely that you would have noticed the speaker looking at the salt, or pointing at it, and from this visual information, you could probably have guessed what he wanted.

But in order to develop any knowledge of German, you would need to start breaking down the incomprehensible stream of speech sounds into possible words (or else you would need a very strong memory to retain the whole phrase!). What the German speaker said was:

Working out the words from a foreign language sound stream can be very challenging. Here's what two English speakers with no knowledge of German thought they heard:

As you can see, they were unable to identify all the separate words, even though the sound system of English is quite close to that of German. In connected speech, there are rarely clear pauses to indicate each word. And if we are not used to the sound patterns of the foreign language, we are unlikely to hear all the elements anyway.

But let's assume you have identified what you think are some words.

How do you find out what they mean? As Klein explains:

Suppose you have really succeeded in identifying certain units in the sound stream of the German speaker, and that you have become aware of two sound entities in particular: [zalts] and [das]. Let us further assume that precisely at the time of saying [das zalts] the speaker had pointed to the salt. Why should [zalts] rather than [das] be associated with the salt on the table? Well, this depends crucially on one's knowledge. An English native speaker might find it plausible that the sound sequence [zalts] means salt. For a speaker of Japanese, however, there is not a trace of a clue like this; what remains is pure guesswork. (Klein, 1986: 59)

To make any sense of foreign language speech, we have to engage in what Klein refers to as 'innumerable cycles of hypothesising', using whatever existing knowledge - knowledge of the world and knowledge of other languages - that we can muster. Arguably teaching can make things easier by isolating and clarifying phrases, but that still leaves the challenge of integrating that knowledge into a language system and using it in communication.

Reflective task 2

1) English data
In comparing the English native speaker and the English non-native speaker, you probably noticed that:

• The range of vocabulary used by the non-native speaker was more limited:
  The native speaker used a variety of expressions to convey the detail - the woman is fat, old, a bun on her head, towards the back of her neck, a hat on her head, with a checkered skirt, tennis shoes, holding a leash; the dog is barking, one of those big dogs, like a greyhound, a saddle, a sweater on his back, a black collar; the man is bald, has a moustache, big nose, wearing a bowtie, with a black umbrella, etc
  
  The non-native speaker's description relied on the key nouns - man, dog, woman, hat, umbrella. There were few more specific nouns and few adjectives.
• Some word choices by the non-native speaker seemed odd or approximate, not what a native speaker would use:
  • the man don't appreciate the dog;
  • because this dog is not quiet (is barking?).
• There was greater fluency and continuity in the native speaker's discourse:
  The language just runs on and on, with the native speaker using phrases like And, I don't know, like, kind of like, to keep the flow going; we have the impression that the language mirrored his thinking. A lot of information was contained in the units or 'sentences' of this speaker's description by means of a lot of adjectives and prepositional phrases, such as with a bun on, on her head, like, towards the back of her head, as well as relative clauses: who's barking, who's wearing, that's got ...
  
  The units of the non-native speaker's speech were much shorter and less complex, consisting mostly of the basic subject + verb + complement type. Pauses were filled with uhh, um and ahh.
• Certain morphological features were missing or overgeneralized in the non-native speaker's speech. (The symbol ø is used to indicate a missing element).
  • lack of the s marker on 3rd person singular, so the 'basic' form of the verb is overgeneralized: lookø, wantø, lookø, knowø;
  • overgeneralization of plural marker s: two mens;
  • overgeneralization of don't to all parts of the verb: the man don't speak.
• The use of some morphological features seemed variable
  • choice of tense in The woman is speaking; the man don't speak seemed arbitrary;
  • sometimes the determiner a was used where required, sometimes not: I see a man, they have ø problem.

2) French, German and Spanish data
You've probably identified similar characteristics in the other descriptions:

• Lexical range is restricted, often with an approximate term, a paraphrase, or an L1 term being used to try to communicate the precise idea:
  • Son chien est fâché et son chien ouvre son bouche (aboie?);
  • El perro umm está gritando al hombre (ladrando?);
  • y está trayendo una umm no recuerdo la palabra, pero umm es para ... para la lluvia .. para no sacar la lluvia;
  • Wir haben eine picture mit zwei Leute;
  • es un bowtie ... la mujer tiene un leash.
• Phrases may be constructed along the lines of the L1:
  • son chien ouvre son bouche (opens his mouth) ;
  • je pense ø la femme est vieux aussi (I think the woman is old as well);
  • ich weiss nicht dieses (I don't know this);
  • Von der links (on the left);
  • en mi opinion (in my opinion)...

In terms of morphology:

• There is variability in tense usage - it seems random:
  • Il y a (present) aussi un homme, un homme avait (imperfect) peur du chien, et le homme porte (present) une....
  • la mujer tiene (present) un umm leash ... La mujer parecía (imperfect) como ella está diciendo (present continuous) algo pero no se...
• There is variability in the use of determiners: sometimes they are missing or the wrong determiner is used; in other cases, the determiner does not 'agree' with the noun it relates to:
  • La femme porte un chapeau, une robe et ø chaussures;
  • die Frau hat einen Hut und nicht in Seite, in Mitte ist einen Hund und rechts ist einen Mann;
  • Pués, hay una mujer con el perro y hay otra hombre.
• Generally, morphological agreement is variable:
  • je pense que la femme est vieux aussi;
  • Der Mann sind in der Nahe der Hund;
  • el perro es suya no y...suyo.

Reflective task 3
Under pressure to communicate your meaning in as few a number of words as possible, you probably chose to concentrate on lexical items (or 'content' words) such as arrive - train - Brighton - tomorrow - 6.30pm leaving out the grammatical items such as I'm - by - in. You no doubt left out the contextualizing phrase As I said yesterday, leaving the reader to fill in the context which you assume he/she knows already. If number of letters had also counted, you would probably have simplified arriving to arrive, deleting the grammatical item ing which carries information about timing, also communicated by the words tomorrow and 6.30pm.

Faced with the demands of the second language acquisition task, second language learners simplify in the same way, concentrating on lexical items: the grammatical 'glue' only gets added gradually.

But why is the grammatical 'glue' important? Without out it, language can be very ambiguous. The message arrive - train - Brighton - tomorrow - 6.30pm could easily mean: I'm arriving by train, but I have to be in Brighton tomorrow at 6.30 pm or I'm arriving by the train that leaves Brighton tomorrow at 6.30 pm or even He/She/They arrive by train in Brighton tomorrow at 6.30pm.

Reflective task 4

1. There are three womans
   Here the plural marker s has been overgeneralized to one of the few words in English where the plural is formed differently.
   
   
2. He get ups early
   Here the 3rd person singular marker s has been applied to what the learner has construed as the verb [get@p]. This is not so much a case of overgeneralization as mis-segmentation of getup, with the result that the learner has applied the right morpheme (s) to the wrong unit.
   
   
3. Il a offri un cadeau à sa mère
   Here the learner has overgeneralized the 'regular' ending for past participles of -ir verbs in French: finir -> j'ai fini, so why not offrir -> j'ai offri? It should of course be offert.
   
   
4. C'est le livre que j'ai lisé l'année dernière
   Here the learner has overgeneralized the standard ('default') form of past participles, which sound similar to the vous form of the present tense: vous chantez - j'ai chanté, vous lisez -> j'ai lisé. Unfortunately, the conjugation of lire does not follow this pattern; its past participle is lu.
   
   
5. Il y a beaucoup de festivaux en France l'été
   Here the learner has overgeneralized the plural ending for nouns ending in - al: cheval -> chevaux, hôpital -> hôpitaux, so why not festival -> festivaux? No reason, it's just festival -> festivals!
   
   
6. C'est un homme vieil
   If the learner had transferred the pattern of their L1 (English), they would have placed the adjective vieil in the correct position! If we take vieux as the 'default' adjective (old), then the learner has managed to make the change for a masculine noun beginning with a vowel. However, they have overgeneralized the 'default' position for adjectives in French, which is after the noun.
   
   
7. Das ist nein kaputt
   The negator is in the right position, it's just the wrong negator, but then German has three. Nein (No) has been overgeneralized to a role where a different negator (nicht) is required.
   
   
8. Nein, ich bist keine müde
   The learner has overgeneralized the 2^nd person singular form of the verb (Du bist...) perhaps because they have heard it again and again in classroom exchanges. They have also overgeneralized the negator used with nouns (kein). But again, the negator appears in the correct position - which is progress!
   
   
9. Da war eine Party und war ich spät nach Hause.
   This learner seems to have picked up on the fact that German often inverts subject and verb, particularly after conjunctions and adverbs (eg here Da). However, they have overgeneralized this feature to a context where it is not used, ie after coordinating conjunctions such as und.

Reflective task 5
You probably attributed the Spanish speaker's I no understand to influence from the speaker's L1 ([yo] no entiendo). However, this explanation cannot hold for German where the L1 form is ich + verb + nicht. Yet Wode (1981) noted that his German L1 daughter consistently produced phrases such as I no drink milk during her acquisition of L2 English in the United States. It seems that learners from all language backgrounds go through similar stages of development in acquiring negation in English.

Reflective task 6
The explanation offered by the two studies is similar. Weinert explains her data thus:

Considering the beginners' status of our learners (Year 1), the high proportion (90%) of postverbally negated sentences containing finite main verbs is surprising. Naturalistic learners distinguish finite main verbs from other verbs, placing the negator in front of finite main verbs... However, there is evidence that these structures when produced by our learners are formulaic, rather than being generated by a learnt system of target-language negative rules. In almost 75% of cases of postverbally negated finite main verbs, subjects insert the adverb gern into the sentence where it is inappropriate or not required. (Weinert, 1987: 92)

In other words, the first year learners were able to produce correct negative sentences because they were reproducing from memory a 'chunk' of language, eg Ich ... nicht gern, which had been exhaustively drilled in class. Similarly, the early correct use of kein was explained in terms of repeated drilling of Ich habe kein X.

As these learners were exposed to more varied uses of nicht and kein, they were not able to integrate the distinction systematically in their internal grammar. Instead, there was considerable variability (randomness, unpredictability) in their selection of kein or nicht. As the data show, their internal grammar tended to favour placing nicht in front of the element to be negated, eg Ich nicht spiele Fussbal. As you may remember (section 2.1.4.1), this placing of the negator in front of the element to be negated seems to be the typical 'basic' pattern in early second language acquisition, produced by both naturalistic and instructed learners of English and of German.

Similarly, although in a different area of grammar and with a different target language, Myles et al comment:

[Our data] indicate that, overall, our learners increasingly resort to V-less utterances over time in order to ask questions [...]. This can be linked to the fact that learners generally rely on chunks less and less as time goes on and as their communicative needs increase. (Myles et al, 1999: 70)

Reflective task 7

1. Comment t'appelles-tu le garçon?
   Here the question Comment t'appelles-tu? has been learned as a chunk and is used simply to ask the question 'What's .... name?', with the subject of the question specified in the following noun phrase le garçon.
   
   
2. Mon petit garçon ... où habites-tu?
   Here the question où habites-tu? has been learned as a chunk and is used simply to ask the question 'Where do/does .... live?' The subject of the question is specified in the first noun phrase. Interestingly, the learner has replaced the determiner le with the determiner mon. This might be evidence of further chunk-type learning, since it is quite possible the learner has memorized phrases beginning with mon/ma + noun.
   
   
3. une famille j'habite un maison
   Here the memorized answer (if we assume classroom drilling of a typical exchange Où habites-tu? J'habite....) forms the chunk, with the subject specified in the first noun phrase tagged on to the chunk.
   
   
4. Où est la gare l'église?
   Here the chunk is the question Où est la gare which has been memorized as a whole (it was the 'presentation phrase' drilled repeatedly in a class on asking directions). The specific reference in this case l'église is then tagged on the end.
   
   
5. What d'you doing, this boy?
   Again, the chunk is What d'you doing, with the specific subject tagged on the end. Interestingly, the chunk here seems to be an amalgam of two chunks What d'you and What' re you which can be phonologically similar in fast, relaxed speech.
   
   
6. Wie alt dein Geburtstag?
   Wie alt must be part of a larger chunk, presumably the question Wie alt bist du? (How old are you)? The learner uses part of the chunk in a similar context to ask not age, but birthday. It is as though Wie alt, presumably known from Wie alt bist du? stands in for a question asking When...
   
   
7. Q: Hast du eine Katze?
   A: Nein, hast du eine Katze.
   Here the learner has simply taken the question and, in order to give a negative answer, has placed the negator nein at the start, and then repeated the question. This repeating back, with some addition at the beginning or end, is found fairly frequently in early interlanguage communication.

Reflective task 8

1) French
According to the Contrastive Analysis Hypothesis:

1. English learners of French will make the error of placing the object pronoun after the verb: Je vois LES.
   Evidence: CAH confirmed ü
   Several researchers have found examples of anglophone learners placing object pronouns after the verb, eg Le chien a mangé les, Il veut les encore (Ervin-Tripp, 1974; and Selinker, Swain and Dumas, 1975, cited in Gass and Selinker, 1994: 63).
2. French learners of English will make the error of placing the object pronoun before the verb: I THEM see.
   Evidence: CAH not confirmed X
   Zobl (1980, cited in Gass and Selinker,1994: 63) was unable to find any evidence of this error in his data from French-speaking learners of English, and none has come to light since. French learners appear not to have problems with putting the object pronoun after the verb in English. The difference between French and English does not result in difficulty in this case.

2) German
According to the Contrastive Analysis Hypothesis:

1. English learners of German will have difficulty pronouncing the final consonant of the German Hand as /t/. They will pronounce it as /d/.
   Evidence: CAH not confirmed X
   According to Eckman (1977, cited in Gass and Selinker, 1994: 97) English learners find it relatively easy to devoice the final consonant, pronouncing Hand with the target-like /t/ at the end.
2. German learners of English will have difficulty pronouncing the final consonant of the English hand as /d/. They will pronounce it as /t/.
   Evidence: CAH confirmed ü
   According to Eckman, German learners have great difficulty in voicing the final consonant, pronouncing hand with L1-influenced /t/, rather than the target-like /d/.

Reflective task 9
According to the Contrastive Analysis Hypothesis, learners assume that the target language corresponds to their L1 pattern, so difference causes difficulty while similarity makes for ease of learning. The sequence of verb marking is identical in all three languages: If + V + past, V + conditional. Thus, in principle, none of the learners in this group should encounter any particular difficulty. However, the reality is somewhat different…

There are unfortunately no data available on French and English learners of Dutch but English and Dutch learners of French (even at advanced level) produce:

Si j'aurais* l'argent, j'achèterais une voiture

while French and Dutch learners of English (at advanced level) produce:

If I would* have the money, I would buy a car

(Source: the author's collection of learner errors and Kellerman, E, 1989).

In this area of grammar, then, learners from all three language backgrounds seem to disregard not only the pattern of their own language but the target language model as well - which frankly seems rather perverse! Why?… Read section 2.1.6.3.

Reflective task 10
Sjoholm and Ringbom's data suggest a strong influence for knowledge of a language that is close to the target language. Then, all other things being equal, an L1 is likely to have more influence than an L2: Swedish L1 learners of English generally did better than Finnish L1 learners who had L2 Swedish. But Sjoholm's and Ringbom's research points out that while having a good knowledge of a close language may be an advantage, it can also be a hindrance: the close language (Swedish) was a more likely source of error than the distant language (Finnish). This again challenges the rather simplistic predictions of the Contrastive Analysis Hypothesis.

Reflective task 13
The results suggest three significant conclusions:

1. Grammar rules seem to be hard to retain and to use accurately, even on a correction task.
   Evidence: in only 46% of cases were the participants, all of whom had been 'taught' the rule on several occasions, able to state an accurate and appropriate rule.
   
   
2. To correct incorrect sentences, you don't have to know the appropriate rule. Evidence: 78% of test items were correctly corrected; accurate rules were only given for 46% of test items. Thus 32% of test items were corrected without recourse to accurate explicit rule understanding. Students must have been using knowledge induced implicitly.
   So far, the obvious conclusion might be 'teaching formal grammar is of little use'. However...
   
   
3. If you do know the appropriate rule, it seems to help accuracy. Evidence: In 97% of those cases where a correct rule was given (46% of total test items), the correction made was accurate.

What seems to be the most sensible conclusion, then, is that formal grammar (usually built up through explicit, deductive, concept learning) is not a straightforward tool in second language learning. It seems to help some students in some cases, but it does not help all students in all cases. The research demonstrates that explicit grammar cannot be necessary to using a second language - otherwise, we would see no accurate corrections without accurate rules - but that it can facilitate accurate use.

The question of how explicit and implicit knowledge work together in second language development is at the forefront of current SLA research (for useful book-length discussions, see Ellis, N C, 1994; Ellis, R, 1998; and Skehan, 1998). This debate was originally triggered in the 1970s and 1980s by attempts by some researchers (in particular, Krashen) to argue that learning a second language relied on the same implicit processes as learning one's first. It is to the notion of 'learning' as 'implicit acquisition' that we now turn our attention.

Reflective task 14
Exercise 1:
You probably arrived at the following answers just by checking for patterns among the data provided:

The rules are as follows:

• nouns ending in -ment are masculine and take le;
• nouns of two or more syllables ending in -age are masculine and take le;
• nouns of one syllable ending in -age are feminine and take la;
• nouns ending in -tion or -sion are feminine and take la;
• nouns ending in -té are feminine and take la;
• both le and la are reduced to l' when the noun starts with a vowel.

Exercise 2:
You can see here how a surface pattern - to be + adjective + to please (+ object) - cannot always be generalized: induction has its limits. The problem here is that easy + to please has a different underlying structure to the other two apparently identical structures. In easy to please, the subject of 'easy' is the object of 'please', since the meaning is in fact: John is easy to be pleased. How can children work out the difference just from hearing the surface structures? Now read section 2.2.3.2.

Reflective task 16
Krashen claims that language teaching methods only have to do two things:

1) supply comprehensible input
This, according to the input hypothesis (Hypothesis 4), will allow acquisition to take place. Nothing else has to be done since acquisition cannot be influenced by explicit learning (Hypotheses 1 and 2) and the acquired grammar will build up on its own according to the 'Natural Order' (Hypothesis 3).

2) ensure a stress-free environment
The only possible blocks to the acquisition process once comprehensible input is available are affective factors (Hypothesis 5). Thus, to minimize these, a stress-free environment is needed and students shouldn't feel pressure to talk (unnecessary, since speaking will 'emerge' on its own).

In the example given here:

• the teacher is using the context to ensure that the learners understand new language (comprehensible input);
• the learners don't have to produce anything in the target language (no emphasis on speaking, a stress-free environment);
• there is no emphasis on conscious learning of language (learning is unnecessary since its value is marginal).

Reflective task 17
The immersion schools set up in officially bilingual Canada were a useful test-bed for Krashen's Input model, providing what Krashen claimed was an optimal environment for second language acquisition.

However, Swain's (1985) research suggests that even after years in this apparently optimal francophone environment, anglophone students' grammatical proficiency in French was significantly different from their native francophone counterparts, who had followed the same curriculum in their native French-speaking schools.

On the positive side, the immersion students' ability to use French in socially acceptable ways was equal to their francophone counterparts. Their ability to organize their French effectively in summaries and narratives was also similar. So there is no doubt that the immersion environment was conducive to the acquisition of fluency in French. However, the differences in grammatical proficiency suggest that comprehensible input is not 'enough' to bring about the development of a complete and accurate grammatical system.

Reflective task 18
Based on the review of research in this area by Ellis, R (1994: 251-65), the kinds of things which NSs do to try to get meaning across to NNSs with limited proficiency are:

• speak slower, with longer pauses;
• use shorter sentences;
• reduce subordinate constructions and increase coordinate constructions;
• use common lexical items repeatedly (little variation of lexical items).

NSs also:

• reduce the amount and simplify the type of information conveyed;
• use more questions;
• focus talk on the here-and-now;
• repeat;
• check understanding from the NNS.

In response to talk from a NNS, they:

• adapt to whatever topics the NNS picks up;
• avoid correcting any errors;
• ask for clarification if understanding breaks down.

As we mentioned, Long (1983) has argued that interactional modifications, such as using more questions, repeating, checking understanding, asking for clarification, are more important in bringing about understanding than linguistic modifications, such as reducing subordinate constructions and shorter sentences, on their own.

You probably guessed correctly that in Pica et al's research, it was Group B (with opportunities for interaction) who were able to complete the comprehension task more quickly and more successfully than Group A who just followed simplified input.

This suggests that interaction between speakers is important for comprehension; in other words, providing comprehensible input is not simply the task of the proficient speaker; learners also need to be active in order to indicate when they don't understand and check when they think they have. There need to be opportunities for negotiation of meaning.

However, while Pica et al's research suggests that opportunities for interaction can lead to better comprehension by non-native speakers, it does not provide evidence that they lead to enhanced acquisition. This kind of evidence has so far been scarce in SLA.

Reflective task 19

What seems to emerge from Richard Schmidt's diary is how the more formal instruction seems to trigger his ability to 'notice' features of the grammar and he seems to be able to use learned knowledge to produce language in conversation:

This week (in class) we were introduced to and drilled on the imperfect... A came over ... I noticed that his speech was full of the imperfect, which I never heard (or understood) before, and during the evening I managed to produce quite a few myself, without hesitating much. Very satisfying!

Now, Schmidt acknowledges that much of his Portuguese development was Krashenian 'acquisition', but his point is that his formal 'learning' seemed to speed up his implicit acquisiton, by facilitating the process of noticing.