A Contrastive Analysis of Conditionals in English and Moroccan Arabic: Potential problems when learning English conditionals by Moroccan EFL learners

： Learning to handle hypothetical situations in a new language is always difficult (Catford, et al., 1974). This rule holds true for Moroccan Arabic (henceforth MA) speakers learning English because grammatical devices in the two languages differ in almost all equivalent situations. For instance, while English verb forms are used to indicate tense in conditional sentences, MA uses them to indicate aspect. Adopting the typology of conditional constructions suggested by Dancygier (1999) and Dancygier & Sweetser (2005), this study provides a contrastive analysis of conditionals in English and MA to predict the possible errors EFL/ESL learners are likely to make while learning English. The analysis shows that the main discrepancy between English conditionals and MA conditionals lies in the verb form used by the two systems. Accordingly, if EFL/ESL learners are influenced by verb form in their L1, they are likely to face some challenges while learning English conditionals. That is, they are likely to use the past tense in the protases of English predictive conditionals and generic conditionals since the perfective form of the verb is used in the protases of these two types in MA. Concerning the protases of English non-predictive conditionals, Moroccan EFL/ESL learners are likely to use either the past tense or the present tense since both the perfective and the imperfective forms of the verb are possible in the protases of MA non-predictive conditionals. However, due to the fact that the perfective form is the prototypical form in the protases of conditionals in MA, EFL/ESL learners are likely to use the past tense more EFL learners apodoses of English conditionals since the prevalent form in the apodoses of MA conditionals is the imperfective.


Introduction
Conditional constructions are among the language aspects that remain challenging for learners regardless of the teaching approaches, methods, and techniques used due to their syntactic and semantic complexities and/or other inherent language-specific characteristics (Celce-Murcia & Larsen-Freeman, 1999). This is due to the fact that learning to handle hypothetical situations in a new language is always difficult as Catford and his colleagues (1974) indicate. Moroccan learners of English are no exception to this generalization because grammatical devices in English and MA differ for almost all equivalent situations. First, verb forms in English and MA will be discussed in this paper. Second, second language acquisition studies will be presented briefly. Finally, on the basis of Dancygier (1999) and Dancygier & Sweetser (2005) classification of conditionals, a contrastive analysis of conditionals in English and MA will be presented and discussed. Possible errors will be highlighted in the contrastive analysis.

Tense and aspect in English and Moroccan Arabic conditionals
In the literature, English is classified as a tense and aspect language (Gaudefroy-Demombynes & Blachère, 1952;Ziadeh & Winder, 1957;Beeston, 1970;Comrie, 1976Comrie, , 1986Declerck, et al., 2006). Comrie defines tense as a "grammaticalized expression of location in time" (Comrie, 1985:9). Declerck, et al. (2006) states that the verb form denotes the location of the situation referred to in time. In other words, tense relates the time of the situation referred to to some other time, usually to the moment of speaking (Comrie, 1976). Tense in English is grammaticalized in verb forms. For example, the past tense is marked by Verb + ed; the present tense is marked by the base form of the verb in the first singular, second singular, and plural person, and Verb + s in the third singular person. Hamm & Bott (2018) identifies two tenses in English, namely past and non-past tense as in (1), (2), (3): (1) James decided to pursue his studies at the faculty of education.
(2) I decide to pursue my studies at the faculty of education.
(3) The teacher will work on his challenges.
In English, verb forms are typically used to signal the time when an action or event occurs or a state holds. Thus in (1), James's decision was taken before the present time; and in (2) the decision is simultaneous with the present time; and (3) states that the teacher's work on his challenges will occur at some time after the present. It is worth to note that the verb forms used to express temporal information may also be used to signal information that is not purely temporal. Consider the following examples: (4) Nick sleeps.
The present tense form in example (4) describes a certain habit of Nick, and in example (5), the present verb form (flies) has a futurate meaning. Therefore, English plainly distinguishes between past and non-past but not between future and non-future. Furthermore, future is excluded as a pure tense by many linguists (Hamm & Bott, 2018), because the future time is reflected in the modal auxiliary verbs as in the following example: (6) She will go fishing in the river.
Tense is contrasted to aspect. Comrie defines aspects as "different ways of viewing the internal temporal constituency of a situation" (Comrie, 1976:3). In the same vein, Declerck, et al. (2006) states that aspect refers to how verb forms are used to express various meanings that the speaker wants to convey by representing the internal temporal structure of a situation. In other words, aspects are ways to view the internal constitution of an actualizing situation. These ways are expressed by different markers on the verb such as suffixes, auxiliaries or a combination of the two, as in the English progressive form. In English, progressive and nonprogressive aspects are the only aspects that are systematically expressed by special verb markers (Declerck, et al., 2006). Consider the following two examples: (7) I'm listening to music. (progressive aspect: the situation is represented as ongoing, i.e. as being in its 'middle') (8) I go to school on foot. (nonprogressive aspect: the reference is to a situation (in this case: a habit) as a whole) Example (7) is marked by the progressive aspect. That is, the situation in this example is represented as ongoing; thus, it is in its middle. By contrast, example (8) is marked by the non-progressive aspect. That is, the situation is represented as a whole. Declerck, et al. (2006) confirms that there is perfective aspect in English. This is the case "when the verb form used reflects the fact that the speaker wants to refer to the actualization of a situation in its entirety, i.e. that he views the situation as if it were a temporally unstructured whole" (Declerck, et al., 2006:30). In other words, the situation is not referred to as having an internal structure (with a beginning, middle and end). For example: (9) I wrote an essay last night.
(10) I will write an essay tomorrow.
In these sentences, the speaker conveys a perfective meaning by using the verb forms wrote and will write. However, Declerck, et al. (2006) prefers not to call them 'perfective verb forms' (as some grammarians do) for the simple reason that non-progressive forms do not always express perfective meaning. For instance, wrote in [they decided to write a letter. Jane dictated] while Mary wrote does not receive a perfective interpretation. In general, English is a tense as well as an aspect language.
Unlike English, MA is an aspect language. There are two basic sets of verb forms, traditionally referred to as aspects and distinguished either as Perfective or Imperfective (Harrell, 1962;Brustad, 2000). The perfective describes a completed event, and relative past time reference. The imperfective describes an incomplete event, and a relative non-past (Harrell, 1962;Brustad, 2000). Harrell (1962) claims that the perfective in MA indicates complete past actions in independent clauses, for example: In this case, the MA perfective is equivalent to the English past tense. When the perfective is used with the copula kan 'to be', it indicates a perfective state in the past as in: (12)  This difference between English and MA as far as tense/aspect is concerned will certainly affect how conditionals are formed in the two languages.
In MA if-conditional sentences, the perfective is used in the if-clauses introduced by indefinite relative pronouns or adverbs, e.g. (16). In such cases, the perfective itself has no time implications at all, and the time of the entire sentence is indicated either by the result clause or by general context. When the perfective is used in the main clause of a conditional sentence in MA, it is usually equivalent to the present or the future in English, e.g. (17) In comparison, the imperfective form in MA indicates various meanings in independent clauses (Harrell, 1962). It is used to indicate immediate future, e.g. (18) General context is the only indication as to the appropriate English translation (Harrell, 1962). It refers to states or to habitual or repetitive actions, never to progressive action with intransitive verbs e.g. (26).
dur be.impf.3s in the office his everyday 'He is in his office every day.'

Second language acquisition studies: the effect of L1 on the acquisition of English conditionals
The influence of L1 on the acquisition of English by L2 learners has been studied extensively (Ellis, 1984;Gass & Varonis, 1994;Schachter, 1974, Ortega, 2009. The influence of L1 on L2 acquisition is argued to be universal. That is to say, all learners are supposed to be influenced by their L1 when learning L2. In this sense, Ortega (2009:48) claims that "previous language knowledge is an important source of influence on L2 acquisition, and this holds universally true of all L2 learners".
Second language studies on conditionals consider that the influence of L1 plays an important role in their acquisition process (Chou, 2000;Lai-chun, 2005;Jung et al., 2005;Kim, 2007;Seo, 2009;Ko, 2013;Tavakoli et al., 2014). All these studies argue that the L1 system of conditionals influences the acquisition of English conditionals by L2 learners.
The influence of L1 on L2 is one of the factors which have been examined by Chou (2000). Chou hypothesizes that evidence for an L1 transfer can be found in the production of conditionals by Chinese learners. He predicts that this effect would appear in three situations, as indicated by Zobl (1982) (cited in Chou (2000): (1) a delay in the restructuring of forms that is necessary for learners to proceed to the next developmental stage; (2) over-production of certain forms; (3) the smallest possible rule change in the passage from one developmental stage to the next one. To test this hypothesis, two groups of participants were recruited for the study: 20 native speakers of English and 36 Chinese speakers.
Results show that there was strong evidence of L1 transfer. Chinese ESL learners prefer to use simple past tense for both past counterfactual conditionals and mixed time reference conditionals. Simple past tense was frequently used in past counterfactual (53% of participants) and mixed-time reference counterfactual conditionals (56%) by Chinese participants. However, it was used less frequently by native speakers of English (21% for mixed-time-reference and 30% for past counterfactual conditional). Chou claims that the high rate of these systematic variations in the production of Chinese participants indicate that Chinese participants tended to prefer the smallest possible rule change in the passage from one developmental stage to the next one. Chou (2000:87) claims that they do that due to the reluctance to make a rule change which "might cause a delay in the restructuring of forms that is necessary for learners to proceed to the next developmental stage, which is another L1 transfer effect".
Another L1 transfer effect is witnessed in the use of will instead of would in the main clauses of the present counterfactual and the mixed-time-reference counterfactual conditional even though they used the right past tense in the if-clauses of the same counterfactual sentences. Chou points out that Chinese speakers use time reference words to determine the time reference of Chinese conditionals. Accordingly, these variations were possibly caused by the use of the adverb now in the main clause of the two counterfactual conditionals. This is supported by the fact that some Chinese participants tended to use the [-past] modal will in the main clauses of the past counterfactual and mixed time reference conditionals when a present temporal reference word now appeared in the same sentence, while they used past tense in the if-clauses.
Likewise, Lai-chun (2005) aims to examine the effect of the mother tongue of the Chinese ESL learners on the acquisition of English conditionals. He hypothesized that negative L1 transfer plays a part in the production of the errors and the over-production of certain forms. To test this hypothesis, a written Chinese-English translation task and a written blank-filling exercise were given to 98 students. Similar to Chou's (2000) study, results in this study show that some participants tended to use will in the main clauses more frequently than in the if-clauses. Therefore, there is an over-production of this form by Chinese ESL/EFL learners. Lai-chun (2005) claims that there is a tendency among Chinese learners of English to use Conditional type 1 with 'will' in the apodosis. This is because Chinese conditionals only have one verb-tense pattern to express different degrees of hypotheticality; therefore, when shifting from Chinese conditionals to English conditionals, there is a tendency for the ESL learners not to change the verb forms to address the different conditional situations. Instead, they prefer to use the verb patterns that are more familiar to them.
In the same line, Ko (2013) points out that L1 influences the acquisition of English conditionals by ESL/EFL learners of English. In his study, Ko aims to check which factor among grammatical complexity, input frequency, and the influence of L1 on L2 can best explain the acquisition of if-conditionals by Korean and Spanish ESL/EFL learners. A total of 130 participants in five groups with 26 each were recruited for the study: native speakers of English (NS), higher-level (KH) and lower-level (KL) Korean-speaking learners as well as higher-level (SH) and lower-level (SL) Spanish-speaking learners. The significance of the effects of L1 (Spanish, Korean vs. English) on the scores of different types of If-conditionals (PG, FP, PC, PPC vs MTR) was tested to examine the influence Korean and Spanish has on the acquisition of if-conditionals by L2 learners. The occurrence of non-target forms was counted and tested by Log Linear Analysis test to examine the effect L1 has on the frequencies of the non-target forms.
Results show that there were no significant differences between the data of the two groups of learners with different L1 backgrounds. That is, the difficulties they found in the different types of if-conditionals between the speakers of Korean and Spanish did not differ significantly: Korean (M: 4.93, SD: 1.87); Spanish (M: 4.99, SD: 1.85). Concerning the results of the non-target forms, a total of 1,135 non-target forms were found in the learners' data. The most frequent type of the non-target forms was the occurrence of the past tense in both the if-clause and the main clause. Statistical analysis based on the frequencies of the six most frequent non-target forms shows that the most commonly found non-target forms in each type of If-conditionals varied significantly (χ2(20) = 489.366, p < .001). For instance, the occurrence of the present tense in the if-clause and the future tense in the main clause was the most frequent non-target forms for the preset generic (PG) if-conditional type, while the past tense in both clauses was the most frequent non-target form for the mixed-time reference (MTR) if-conditional type. Concerning the differences between Korean and Spanish participants, the two groups of learners differed significantly (χ2(5) = 23.949, p < .001) in terms of the types of the non-target forms they produced. While the most frequent non-target form produced by Korean and Spanish learners was the past tense in both the if-clause and the main clause, modal + the past tense in the if-clause and the past tense in the main clause were more frequently used by Spanish learners (13.95%) than by Korean learners (6.52%). Ko (2013) also notes that Spanish learners used modals (33.51%) more than Korean learners (9.09%).
The differences between the most frequent forms found in the results of the participants with different L1 backgrounds can be interpreted by the influence of L1 on L2. That is, in spite of the fact that both "English and Spanish conditionals use tense shifts to express higher level of hypotheticality in the case of the present counterfactual (PC), past counterfactual (PPC) and mixed-time reference (MTR) types of conditionals, they differ in the choice of verb forms" (Ko, 2013:170). In other words, as Ko explains, in Spanish the use of the PC, PPC and MTR types of conditionals involve the use of the subjunctive form of verbs in the if-clause, while in English they do not. The subjunctive meaning expressed by the use of modals along with the verb in the if-clause in Spanish might have influenced Spanish learners' use of conditionals in English. Ko states that the past and the past participle which are used in the if-clauses of PC, PPC and MTR types of English conditionals are in fact present and past subjunctives, just as in Spanish conditionals. The problem is that they look like past and past perfect indicatives in English. This fact justifies the high frequency of the past tense non-target forms in the past conditionals and mixed time reference conditionals among Spanish learners. Conversely, Korean learners of English rarely produce non-target forms with modals in English conditional if-clauses probably because Korean does not have a subjunctive form in conditionals.
In another study, Katip and Gampper (2016) aim to find out which conditional verb forms are most difficult for Thai secondary school students to produce, and what errors would be made by those students in written and spoken English. A gap-filling task was given to 68 students to examine their conditional errors in written English. They were asked to produce grammatically correct verb forms in both the if-clause and the main clause of the target if-conditional types. A spoken task was given to the same group to explore the errors made by 20 students in the first task.
Results of the gap-filling task show that the participants were more likely to have trouble producing grammatically correct verb forms of the past counterfactual and present counterfactual conditionals than the other if-conditional types. The following table shows the overall number of English if-conditional errors found in the gap-filling task: This table indicates that two-thirds of the errors made by the participants were in present counterfactual and past counterfactual conditionals. Accordingly, these two types seem to present a big challenge for Thai learners of English in comparison with the other conditional types.
As for the type of errors found, the results show that the most common errors in both the if-clause and main clause of the factual conditional type were derived from the misuse of subject-verb agreement (96 tokens/26.67%), along with future simple tense (93 tokens/25.83%). The participants in the study applied wrongly these two grammatical points more frequently than others (present participle (45 tokens/12.50%), past simple tense (43 tokens/11.94%), misinformation (38 tokens/10.56%). The most frequent errors found in the future predictive conditional type were the future simple in the if-clause (33 tokens/14.16%) and the present simple and the past simple in the main clause (43 tokens each/18.45%). Among the errors found were the present participle and past participle which together accounted for 19.75 percent of the total number of errors. The subject-verb agreement which was caused by the incomplete use of a grammatical rule in present simple tense was still ranked in the top five conditional errors. As for the present counterfactual conditional type, the most common errors were the present simple tense (223 tokens/45.14%) and future simple tense (166 tokens/33.60%), respectively. Concerning the past counterfactual conditionals, the most common errors were involved with misusing the future simple tense (149 tokens/32.04%), present simple tense (136 tokens/29.25%), and past simple tense (105 tokens/22.58%), respectively. Katip and Gampper (2016) point out that the gap-filling task revealed interesting and unexpected findings. First, the misuse of present simple tense was found in the first rank of future predictive and present counterfactual. Participants misused the present simple tense in the main clause of the future predictive conditional. In addition, participants misused this grammatical structure in the counterfactual conditionals even though this verb tense has nothing to do with the grammatical features of these conditional types. Second, except for the past counterfactual, all the conditional types shared the second most occurred error-the misuse of future simple tense in the if-clause, in the production of conditional verb patterns in written English. Based on these results, Katip and Gampper (2016) conclude that the English conditional types investigated were influenced by the over-production of the verb pattern 'modal + verb', particularly 'will + verb'. That is, as Chou (2000) argues, L2 learners tend to show slight changes in the rule of English if-conditionals from one developmental stage to the next one. For instance, in Katip & Gampper's (2016) study, findings show that present simple tense was still used by the majority of participants even after they have acquired conditionals from the future predictive to present counterfactual and to past counterfactual. Third, the past simple tense structure was among the top five errors of all conditional types in written English. These findings can be explained by Brown's Cumulative Complexity (1973) andO'Grady Development Law (1997). That is, despite the fact that the past counterfactual structures have both [+past] and [+perfect] grammatical features, nearly all of the participants employed only the past simple form when producing both clauses of this conditional type. This means that Thai participants could not acquire the past counterfactual conditional since it has the highest number of grammatical features and is the most grammatically complex noun among the other three types studied (Chou, 2000;Ko, 2013;Lai-chun, 2005).
The results of the spoken task revealed that the most common error was the misuse of future simple tense in the students' English utterances. The students mostly used the construction 'will + verb' in the main clause of the target conditional types. In addition, the incomplete use of the grammatical relationship between subject and verb occurred in their English utterances as well. Put it differently, many students could not manage to make a verb agree with its subject, particularly a singular subject. Accordingly, the findings of the gap-filling task mentioned in the previous paragraph apply also to the findings of the speaking task.
In general, it is shown by all the reviewed studies that L1 influences the acquisition process of L2 conditionals by ESL/EFL learners. This is done either through the overproduction of some forms like the past tense in Spanish and Chinese, or the use of the smallest possible rule change from one developmental stage to the next one. The next section will compare between Moroccan Arabic and English conditionals to predict some potential problems Moroccan learners of English may face when learning English conditionals.

A contrastive analysis between Moroccan Arabic and English in light of Dancygier (1999) and Dancygier & Sweetser (2005) classification
Following the typology provided by Dancygier (1999) and Dancygier & Sweetser (2005), this study will provide a contrastive analysis between MA and English conditionals. The purpose of this contrastive analysis is to predict some potential problems Moroccan learners of English may face when learning English conditionals. Dancygier (1999) and Dancygier & Sweetser (2005) classify conditionals into three main categories and ten sub-types. The following contrastive analysis of conditionals in MA and English will follow this classification.

Predictive conditionals
Predictive conditionals are marked with backshift and a regular pattern of verb forms (Dancygier, 1999;Dancygier & Sweetser, 2005). Verb forms in this category are linked to the degree of unassertability of the proposition expressed in the conditional construction. The relationship between the protasis and the apodosis is marked by causality or sequentiality. Using mental space theory terminology, the speaker in predictive conditionals sets up mental spaces wherein he imagines some future scenarios to make some decisions or take some actions. In this way, the speaker in such constructions engages in prediction (Dancygier & Sweetser, 2005). The predictive conditionals category is subdivided into two sub-types, namely, non-distanced predictive conditionals and distanced-predictive conditionals. The latter is further subdivided into two other types: weak-hypothetical predictive conditionals and strong-hypothetical conditionals.

Non-distanced predictive conditionals
In non-distanced predictive conditionals, the speaker does not distance himself from the commitment of the fulfilment of the condition expressed in the conditional construction (Dancygier, 1999). In this sub-type, English uses if with present tense in the protasis and will + v in the apodosis, as in the following example: In comparison, MA has ila (if) in non-distanced predictive conditionals. In this type, ila is used with the perfective form of the verb in the protasis and ɤadi (will) with the imperfective form of the verb in the apodosis, as in (28) below. ila is also used with the imperfective form of the verb in the protasis and the apodosis as in (28)  In example (28), the verb in the protasis is in the perfective, which often corresponds to the past tense in English. For this reason, Moroccan EFL learners may use past tense in the protasis of English non-distanced predictive conditionals if they are influenced by the equivalent verb form in their L1. Consider the following sentence: (29) If we go on time, we will catch the train.
The equivalence of this sentence in MA is: if go.perf.1p in the time, will catch.impf.1p the train 'If we go on time, we will catch the train.' Since the verb of the protasis is in the perfective form, Moroccan EFL learners may use went instead of go. Concerning the apodosis of sentence (29), Moroccan EFL learners may not face any challenges. This is because the equivalence of the MA word ɤadi is the modal will in English. Moreover, the equivalence of the imperfective verb nʃǝddu is the present verb catch in English. Oppositely, MA learners may not face any problems in the protasis of the following sentence which belongs to non-distanced predictive conditionals: (31) If you go, I'll go.
The equivalence of this sentence in MA is: if go.impf.2s, go.impf.1s 'If you go, I'll go.' The verb of the protasis is in the imperfective form, which often corresponds to the present tense/form in English. However, they may face a problem in the apodosis. If they are influenced by their L1, they may avoid the use of the modal will since the imperfective form of the verb is used without the word ɤadi "will".

Distanced predictive conditionals
In distanced predictive conditionals, the speaker distances himself from the commitment of the fulfilment of the condition expressed in the conditional construction (Dancygier & Sweetser, 2005). As noted by Dancygier & Sweetser, distanced predictive conditionals in English are sub-classified into two sub-types: (a) weak hypothetical conditionals, and (b) strong hypothetical conditionals. Weak hypothetical conditionals indicate that the prediction is strong on the scale of assertability, e.g.: (33) If he went, I would go.
In English, they are characterized by the use of past tense in the protasis and would + v in the apodosis. Strong hypothetical conditionals indicate that the prediction is low on the scale of assertability, e.g: (34) If he had gone, I would have gone.
In English, they are characterized by the use of past perfect tense in the protasis and would + have + past participle in the apodosis.
In comparison, MA generally uses kun (if) in distanced predictive conditionals. kun is used with the perfective form of the verb in the protasis and the imperfective form in the apodosis as in (35) below. If the verb in the apodosis is in the perfective form, it is usually preceded by kun, as in (36)  Examples (35) and (36) suggest that in MA both weak-distanced predictive conditionals and strong-distanced predictive conditionals do exist as in English. However, they do not behave in the same way they do in English. That is, weak-distanced predictive conditionals use the past tense in the protasis and would + the base form of the verb in the apodosis while the time actually referred to in the sentence is the present. Strong-distanced predictive conditionals use the past perfect tense in the protasis and would + have + past participle in the apodosis and the time referred to in the sentence is in the past. By contrast, in MA, the weak-distanced predictive conditionals use the perfective in the protasis and the imperfective in the apodosis, in which the time actually referred to the present.
Strong-distanced predictive conditionals use the perfective both in the protasis and the apodosis with a compulsory kun before the perfective verb used in the apodosis, in which the time actually referred to the past. Therefore, example (35) represents weak-distanced predictive conditionals and example (36) represents strong-distanced predictive conditionals in MA. To differentiate between the two types, MA uses kun in the apodosis of strong distanced-predictive conditionals.
The equivalence of sentence (35) in English is sentence (37) below: (37) If he went, I go.
Since the verb of the protasis is in the perfective form in sentence (35), Moroccan EFL learners are likely to use the right form of the verb in English weak-distanced predictive conditionals. However, they may drop the modal would from the apodosis of this type of conditionals in English because there is no equivalence of such word in MA weak-distanced predictive conditionals.
The equivalence of sentence (36) in English is sentence (38) below: (38) If he went, I went.
Accordingly, MA EFL learners of English may have difficulties in learning strong-distanced predictive conditionals because the verb form used in the protasis and the apodosis of this sub-type in MA is the perfective. If they are influenced by verb form in their L1, they may use the past tense instead of the past perfect in the protasis in English strong-distanced predictive conditionals. In addition, they may use the past tense instead of would + have + past participle in the apodosis of this type.

Non-predictive conditionals
The second category is the category of non-predictive conditionals. As noted by Dancygier (1999), this category does not require any form of backshift or any regular pattern of verb forms. In terms of mental space theory, the speaker in non-predictive conditionals is not involved in setting up alternative spaces as he does in predictive conditionals. In this category, the speaker uses the if-clause to specify the mental space background against which the proposition in the main clause is made. This category is divided into six sub-types, namely, epistemic conditionals, speech act conditionals, meta-linguistic conditionals, meta-metaphoric conditionals, elliptical conditionals, and meta-spatial conditionals.
4.2.1 Epistemic conditionals Dancygier (1990Dancygier ( , 1993 describes epistemic conditionals as conditional constructions in which the protasis presents a premise and the apodosis the conclusion inferred from the premise. In this respect, Sweetser (1990) claims that the knowledge of p in this type of conditionals is a sufficient condition for concluding q. In English, epistemic conditionals do not show any regular patterns of verb forms. Thus, we may find different verb forms in both the protasis and the apodosis. To illustrate, consider the following examples: (39) If he left at nine, he found Adam in the house.
(40) If they leave Casablanca now, they must be here at 10 p.m.
In terms of meaning, both sentences refer to the fact that if the condition expressed in the if-clause is fulfilled, it is enough to conclude the proposition expressed in the apodosis. However, the two sentences refer to two different tenses: sentence (39) refers to the past, while sentence (40) refers to the future.
By comparison, epistemic conditionals in MA behave differently from their counterparts in English. While epistemic conditionals introduced by the conjunction ila are marked by not showing any regular patterns of verb forms in the protasis, they show regular patterns of verb forms in the apodosis. Thus, in MA ila is used with either the perfective or the imperfective form of the verb in the protasis and the perfective form of the verb in the apodosis. Consider the following equivalents of examples (39) and (40) While the English sentences in (39) and (40) use two different verb forms in the protasis (sentence (39) uses the past and sentence (40) uses the present), their equivalents in MA use the same form (sentences (41) and (42) use the perfective form of the verb). As for the apodosis, the past tense is used in sentence (39), while the modal must + the base form of the verb is used in sentence (40). In MA, the imperfective is used in sentence (41), while xas + the imperfective is used in sentence (42). The problem raised by the English examples is that, while (39) refers to the past, its equivalent in MA (sentence (41)) refers to the future. Such problem is not raised by sentences (40) and (42). This is due to the fact that in sentence (40), the verb form indicates the present tense. In sentence (42), the adverb daba 'now' refers to the present time. Thus, in MA epistemic conditionals, the time of the entire sentence is indicated either by adverbs or by the general context.
Since epistemic conditionals require different verb forms both in the protasis and the apodosis, MA EFL learners may probably have difficulties in learning English epistemic conditionals. If they are influenced by verb forms in their L1, they are likely to use past tense in the protasis of epistemic conditionals. Accordingly, MA EFL learners would tend to use the correct form of the verb leave (i.e. left) in the protases of sentences (39) and (40). This is due to the fact that the equivalent of the perfective verb xrǝʒ used in the protases of sentences (41) and (42) is left.
In the apodosis of epistemic conditionals, MA EFL learners are likely to provide the correct form of the verb in sentence (40). This is due to the fact that the equivalents of the modal must and the verb be are xas and ikun respectively. However, they may not generate the correct verb form in sentence (39). This is because the equivalent of the verb form found in the apodosis of sentence (39) is lqa in MA. Actually, this is not the verb form used in MA as an equivalent of found in English as shown in sentence (41). The verb form used in the protasis of sentence (41) is the imperfective ilqa. The equivalent of ilqa in English is the present tense finds. Thus, instead of using found in the apodosis of sentence (39), MA EFL learners may use finds if they are influenced by their L1.

Speech act conditionals
Speech act conditionals are conditionals in which the protases are said to guarantee a successful performance of the speech act in the apodoses (Dancygier, 1999). In English, speech act conditionals hardly ever allow distanced verbs. Consider the following example: However, there are no other restrictions on the verb forms used in both clauses of this type of conditionals.
In MA, however, only the particle ila is possible in speech act conditionals. The particle kun cannot be used in this conditional type, probably because we cannot make an offer, a request, and the like using unreal situations. The protasis in speech act conditionals does not show any regular pattern of verb forms. Thus, both the perfective and the imperfective verb forms are possible in the protasis. However, only the imperfective form of the verb indicating the imperative mood can occur in the apodosis. In sentence (45), the auxiliary verb ma-kan (be.imper.neg, 'wasn't') in the protasis is in the perfective form. In sentence (46), the auxiliary verb ma-kayan 'isn't' in the protasis is in the imperfective form. The MA sentences in (45) and (46) are the equivalents of the English sentence in (43). That is, the affirmative sentence of the if-clause in English is rendered into a negative sentence with two possible verb forms in the if-clause in MA. Concerning the main clause, the future tense is used in English as in (43), while in MA the verb form used is usually in the imperfective indicating the imperative mood as in (45) and (46).
MA EFL learners may probably have difficulties in learning English speech act conditionals if they interpret the perfective form of the auxiliary verb kan as referring to the past. Therefore, they may use:

Meta-linguistic conditionals
Meta-linguistic conditionals, known for their commenting on assertability, are object to a part of the previous utterance, and pertain to the same range of phenomena, as explained by Dancygier (1999) and Dancygier & Sweetser (2005). In other words, the speaker in meta-linguistic conditionals comments on the choice of the linguistic expression used in the main clause. In English, when meta-linguistic conditionals are used to indicate prediction, distanced verb forms are possible, as in (51) below. By contrast, when this type is used to indicate non-prediction, verb forms are unconnected with each other and cannot take distance-marking forms, as in (52) below: (51) If she was driving on the left side, she would be English.
(52) Grandfather is feeling ecstatic, if that's an appropriate expression.
In comparison, in MA, ila and kun are both possible used in meta-linguistic conditionals. The difference between the two lies in the attitude towards the proposition expressed in the protasis, as in (53) and (54)  The protases of sentences (53) and (54) are introduced by the particle ila. The speaker in these two sentences does not distance himself from the commitment of the fulfilment of the condition expressed in the conditional construction. That is to say, it is very likely that the girl mentioned in sentence (53) is driving on the left. Likewise, it is possible to describe the speaker's grandfather as feeling ecstatic in (54). By contrast, the protasis in sentence (55) is introduced by the particle kun. The speaker in this sentence distances himself from the commitment of the fulfilment of the condition expressed in the conditional construction. In other words, it is very unlikely to describe the mentioned person in the protasis of sentence (55) as crazy.
Concerning verb forms used in meta-linguistic conditionals in MA, both the perfective and the imperfective forms are possible in the protasis as in (53) and (54). Therefore, MA EFL learners may not have a problem using the protasis of a sentence like (51) because the equivalent of the progressive form was driving used in the protasis of sentence (51) is the verb form kant katsug used in the protasis of sentence (53). However, they may have a problem using the protasis of a sentence like (52) because the equivalent of the English verb form is feeling in MA is kan ħas. Actually, this is not the right form used in MA. The MA verb form used as equivalent to the English present progressive form is feeling is the perfective form of the auxiliary verb rah 'was'. So, there are two problems that may stand as a challenge to MA EFL learners in learning the protasis of an English meta-linguistic sentence like (52). The first one is the dropping of the verb feel and the second one is the use of the perfective form of the auxiliary verb rah as shown in example (54).
Likewise, in the apodosis of meta-linguistic conditionals both verb forms are possible. However, contrary to the apodoses of sentences like (51) and (52), MA EFL learners will probably face a problem in the apodosis of a sentence like (51) while they probably will not have a problem in the apodosis of a sentence like (52). This is due to the fact that the equivalent of the MA word ɤadi used in the apodosis of sentence (53) (sentence (53) is the equivalent of sentence (51)) is the modal will; however, would is the one used in the apodosis of sentence (51) as the equivalent of the MA word ɤadi. Therefore, MA EFL learners are likely to use the modal will instead of would in the apodosis of an English meta-linguistic sentence like (51). Conversely, MA learners may not have a problem in the apodosis of a sentence like (52). This is because the verb form used in the apodosis of sentence (54) (the MA counterpart of (52)) is the perfective. As is mentioned above, the counterpart of the MA perfective is the present tense in English. Thus, MA EFL learners may generate the right form of the verb in the apodosis of an English meta-linguistic conditional sentence like (52).

Meta-metaphorical conditionals
Meta-metaphorical conditionals resemble epistemic conditionals in that inference is required on the part of the hearer in both conditional types. In meta-metaphorical conditionals, metaphorical mappings are transfers of inference from source to target (Dancygier, 1999;Dancygier & Sweetser, 2005). In English, meta-metaphorical conditionals have no restrictions on verb forms. Thus, no regular patterns of verb forms are required in this type, as Dancygier (1999) and Dancygier and Sweetser (2005) argue. To illustrate consider the following examples: (56) If you are a minister, I am a king.
(57) If you played football very well, I was Ronaldo.
Meta-metaphorical conditionals are usually introduced by the particle ila in MA. Both the perfective and the imperfective forms are possible in the protasis and the apodosis. In (58) and (59), the verb in the protasis is in the perfective form. However, the protasis of sentence (58) has a present time reference indicated by the use of the copula rani 'I am', while the protasis of sentence (59) has a past time reference indicated by the use of the perfective form of the copula kunt 'I was' in the apodosis. Likewise, the copula is used in the apodosis of the (58). It is possible to do without the copula as well. In the apodosis of (59), the copula is necessary.
MA EFL learners are likely to have problems in learning the protasis of meta-linguistic conditional sentences like (56). They are likely to use past tense in the protasis of these sentences. This is due to the fact that the verb form used in the protasis of a MA meta-metaphorical sentence like (58) (the counter part of sentence (56) in English) is the perfective (i.e. kunt "you were") which corresponds to the past tense in English. However, this is not the case in English. As presented above, the equivalent of the MA meta-metaphorical conditional sentence in (58) is the English meta-metaphorical conditional sentence in (56). As can be noticed, the equivalent of the perfective verb kunt in the protasis of sentence (58) is the present tense of verb to be (i.e. you are) in the protasis of sentence (56). Concerning the apodosis of a meta-metaphorical conditional sentence like (56), it is expected to be learnt easily by MA EFL learners. They are likely to use the right verb form (i.e. present tense "I am") in the apodosis of a sentence like (56) because the verb form used in the apodosis of the MA equivalent of sentence (56) is the imperfective form of the auxiliary verb rani "I am" (see example (58) above) which corresponds to the present tense in English.
By contrast, MA EFL learners may not have any problem in learning meta-metaphorical conditional sentences like (57). This is because the verbs used in the protasis and the apodosis of the counterpart of this sentence in MA are in the perfective form which corresponds to the past tense in English. As can be seen in sentence (59), the perfective verb used in the protasis is lʕǝbti "played" which corresponds to the past tense verb played used in sentence (57). The perfective verb used in the apodosis is the auxiliary verb kunt 'I was' in (59) which corresponds to the past tense auxiliary verb was used in sentence (57).

Elliptical conditionals
Elliptical conditionals are conditional constructions that appear to lack elements that can be recovered from context (Dancygier & Sweetser, 2005). Elliptical if-clauses (and sometimes also main clauses) are interpreted exactly as their full forms and omission is employed simply to avoid repetition; that is, for stylistic reasons (Dancygier, 1999).
Unlike English, MA does not have elliptical conditionals. There are two types of constructions in MA which can stand as equivalents to elliptical conditionals: verbal constructions and non-verbal constructions. In verbal constructions, the verb and the participle carry necessarily agreement markers, and that, in the absence of subjects, these markers refer to the person, gender and number of the deleted subjects. Therefore, the recovery of this information is less dependent on context in English (as in (60)) than in MA (as in (61a)). Non-verbal constructions in MA can express meaning without use of the verb as in (61b). Sentences (61a) and (61b) are the MA equivalents of the English sentence in (60). In (60), the missing elements are the subject and the verb in the if-clause. In (61a), there are no missing elements in the conditional sentence. In (61b), the if-clause is a non-verbal construction.
MA EFL learners may probably have difficulties in learning the protasis of English elliptical conditionals. They are likely to think that elliptical conditionals are wrong since they seem to be missing something crucial in the sentence. For instance, they may think that the protasis of sentence (60) is missing a subject and a verb in the present tense. Therefore, instead of saying: (62) if not a jackass They are likely to say:

s not a jackass
Contrarily, they may not have any difficulties in learning the apodosis of English elliptical conditionals. As has been presented in sentences (60) and (61a) and (61b), the equivalent of verb to be in present tense is used in the apodosis of sentence (60) is the MA auxiliary verb rah used in the apodoses of sentences (61a) and (61b).

Meta-spatial conditionals
Meta-spatial conditionals are conditionals in which the speaker relates the proposition in the protasis to the proposition in the apodosis to conclude that the proposition made in the protasis is sarcastic, illogical, or the like (Dancygier & Sweetser, 2005). In English, there is no restriction on the verb forms used in the protasis and the apodosis, as claimed by Dancygier (1999) and Dancygier & Sweetser (2005). Consider the following examples: (64) If your house is bigger than ours, which is bigger: Hicham's apartment or Anas's?
(65) If your house was bigger than ours, Hicham's apartment would be bigger than Anas's.
In sentence (64), the verb to be is in the present tense, and the sentence refers to the present time. In this sentence, the speaker is asking the addressee to reconsider his conclusion toward the size of the two houses. In sentence (65), the verb to be is in the past tense; however, the sentence refers to the present time. In this sentence, the speaker tells the addressee that his conclusion (the addressee's house is bigger than the speaker's) is irrational. He does that by comparing two other apartments which both the speaker and the addressee know (Hicham's apartment and Anas's apartment).
Both conjunctions ila and kun are possible in this type of conditionals in MA. Both perfective and imperfective forms of the verb are possible, as well. To illustrate, consider the following example: The use of ila indicates that the speaker's judgment concerning the irrationality of the addressee is not strong. In example (66), the speaker invites the hearer to reconsider his statement. However, the use of kun indicates that the speaker's judgment concerning the irrationality of the speaker is very strong. In example (67), the speaker states it clearly that the hearer's judgment is irrational by providing a comparison which both the speaker and the hearer know that it is not true; i.e. Anas's apartment is bigger than Hicham's. The nature of the relationship between the first and the second proposition, however, remains unclear and can only be inferred from the background knowledge shared by the speaker and the hearer.
MA EFL learners may probably have difficulties in learning the protasis of an English meta-spatial conditional sentence like (64). The verb to be used in the protasis of sentence (64) is in the present tense (i.e. is); however, its equivalent in MA is the perfective form kan used in the protasis of sentence (66) which corresponds to the past tense of verb to be (i.e. was) in English. Thus, MA EFL learners may wrongly use was instead of is in the protasis of a sentence like (64) if they are influenced by their L1 conditional system. MA EFL learners may not have any difficulties in learning the apodosis of a meat-spatial conditional like sentence (64). This is because the equivalent of the nonverbal construction ʃkun lli kbǝr used in the apodosis of sentence (66) is the English structure which is bigger used in the apodosis of sentence (64). Therefore, they are very likely to use the right structure is this context. Contrary to sentence (64), MA EFL learners may have difficulties in learning the apodosis of a meta-spatial conditional sentence like (65) while they may correctly utter the protasis. In other words, since the verb of the protasis is in the perfective form kant in sentence (67) which corresponds to the past tense was in English, Moroccan EFL learners are likely to use the right form of the verb in an English meta-spatial conditional sentence like (65). However, they may drop the modal would from the apodosis of this sentence in English because there is no equivalence of such word in MA.

Generic conditionals
Generic conditionals describe a predictive relationship between "a state of affairs in P and the causally dependent state of affairs in Q over a generic class of mental spaces. For any given mental space, if P is known to obtain, then the eventuality with respect to Q will be predictable" ① (Dancygier, 1999:95). In English, generic conditionals canonically take the form "IF P-pres, Q-pres" as in (68)  Simple past forms are possible with generic conditionals as in (69)  In conditional sentences such as (70a), the if-clause fulfils the requirement for the background clause of a predictive construction and simultaneously manifests the form usually used in MA for generic event reference. Generic conditionals in MA refer to no time. The generic category is then independent of time. This is proved by the fact that if we use the adverbs lbarǝħ 'yesterday', daba 'now', or ɤǝdda 'tomorrow' in sentence (70a), it becomes ungrammatical. kun is not possible in generic conditionals because generic meaning cannot be expressed by a hypothetical particle which result in the ungrammaticality of sentence (70b). Concerning verb forms, MA generic conditionals show regular patterns of verb forms. Thus, the perfective is always used in the protasis and the imperfective with the durative ka-in the apodosis to indicate the fulfilment of the condition at all times.
It has been mentioned in section (2) that the perfective in MA indicates complete past actions in independent clauses (Harrell, 1962). Thus, MA EFL learners are likely to have difficulties in learning the protasis of English generic conditionals. They may use past tense in the protasis instead of present tense in English generic conditionals. For instance, in the protasis of sentence (68), MA EFL learners are likely to use the past tense played instead of the present tense plays if they are influenced by their L1. This is because the equivalent of the present tense verb plays in English is the imperfective verb ilʕǝb 'plays' in MA. However, as can be seen in sentence (70a), the equivalent of the present tense plays is the perfective verb lʕǝb 'played'. On the other hand, they may not have difficulties in learning the apodosis of this type in English since the verb form used in MA in the apodosis of generic conditionals is the durative which equals the present tense in English. For instance, the equivalent of the present tense wins in the apodosis of the English sentence in (68) is the imperfective verb irbǝħ 'wins' in the apodosis of the MA generic conditional sentence in (70a).

Implications of the study
Students in Moroccan schools are over-exposed to some if-conditional types rather than others. As a matter of fact, Moroccan students are exposed to the conditionals that belong to the classical typology suggested by Comrie (1986). The other conditionals are neglected to the extent that students cannot utter a conditional sentence out of the types that belong to the classical typology. Consequently, "students being equipped with limited structures might consider these the only 'correct' ways of using conditionals and all untaught structures colloquial or 'substandard' usage" (Lai-chun, 2005:54). Therefore, among the main reasons for Moroccan students' difficulties with English conditionals is possibly teacher interference in addition to L1 interference. This idea is supported by Maule (1988) who claims that among the reasons for ESL students' difficulties with English conditionals is teacher interference in addition to L1 interference. This study introduces curriculum designers and teachers to the other types of English conditionals. In this way, all the types of English conditionals will be taken into consideration when designing student programs. It is hard to introduce all conditional types in the program, but as suggested by Fulcher (1991), what to be included in the syllabus should be determined by students' purpose and need to learn particular conditional forms. This study also highlights the possible errors MA EFL learners may make when learning the different types of English conditionals. Accordingly, teachers and curriculum designers are invited to take these errors into account when introducing English conditionals to MA EFL learners. The best way to do that is to highlight the similarities and differences between the two conditional systems to help MA EFL learners avoid those errors.

Conclusion
Although English and MA belong to two different language families, for English is a member of Indo-European language family and MA belongs to Afro-Asiatic language family, there are some similarities between English and MA in conditional constructions in both semantic and syntactic aspects, yet their differences outweigh their similarities. Through this contrastive study, if Moroccan EFL/ESL learners are influenced by their L1, they are likely to face some challenges while learning English conditionals especially at the level of verb forms. Consequently, Moroccan EFL learners may overuse the past simple in their production of English conditionals. Therefore, this study supports theoretically the findings of second language acquisition studies mentioned in section 3. It is hoped that this study can help English learners in Morocco gain a better insight into peculiarities of their mother language and English in conditional constructions so as to avoid mistakes in English writing or daily communications and to improve the English teaching and translation from English into MA and vice versa.