Conditional sentences are a big obstacle to overcome for teachers and students of English as a second or foreign language (ESL/EFL). Conditionals are linguistically and cognitively complex structures that express a variety of meanings, are realized through a variety of forms, and are used for a variety of discourse functions.
    
The complexity of expressing conditional sentences lies in the dependence of one circumstance on the occurrence of another. The variety of possible meanings includes areas of cognitive reasoning, logical argument, psychological intent and desirability, and semantic nuances associated with real, counterfactual, or hypothetical events contingent on, inferred from, or caused by one or more of these events.
    
There is no simple answer to the question of how conditional structures should be taught. Much research has been done on conditionals and certain proposals have been made, but for the practical needs of teachers and students in the classroom no easily-explained or easily-learned solution has been forthcoming.
    
This paper is an attempt to provide a workable shortcut for teachers and students dealing with understanding and producing the numerous types of conditionals found in daily life. This paper first examines relevant research on the difficulties inherent in the teaching and learning of conditionals, then proposes a simple model that goes beyond the traditional way of introducing and practicing these troublesome forms.
    
The sections on relevant research include the main difficulties related to the many types of conditionals; their frequency, number, and usage; their grammatical structures; and some areas of controversy. The proposal section outlines a means of simplifying and combining conditional categories by focusing on (a) time-tense relationships and backshifting¹ and (b) the relationship of "hope" and "wish" sentences to "if" conditionals. A second paper containing detailed lesson plans based on the proposed model will be forthcoming.

Celce-Murcia and Larsen-Freeman (1999) cite a survey conducted by Covitt (1976) that found that conditionals ranked fifth (behind articles, prepositions, phrasal verbs, and verbals) among the most serious teaching problems encountered by ESL teachers in the Los Angeles area. The main difficulties lie in the following aspects:

The form of conditionals is different from other structures in English because they have two clauses: a main clause and a subordinate clause. The subordinate clause typically begins with the adverbial subordinator "if." Quirk and Greenbaum (1973) state that conditional sentences express the dependence of one set of circumstances (i.e., the result clause) on another (i.e., the "if" clause). In addition, the ordering of the two clauses can be reversed in most cases.
    
Variations of the above forms also exist. For example, when the "if" clause occurs in initial position, "then" can be added before the main clause (e.g., "If it rains, then I will take my umbrella"). In hypothetical conditionals with initial "if" clauses containing certain auxiliary verbs (e.g., "had" and "should"), the initial "if" can be deleted provided there is a subject/operator inversion.

Halliday and Hasan (1976) mention that certain pro-forms (e.g., "so" and "not") can be used to replace the entire conditional clause following "if."  

Various other conjunctions and phrases such as "only if," "unless," "even though," "even if," "whether or not," "providing/provided that," "on condition that," "as long as," and "on the understanding that" can also introduce conditional clauses. The main focus of this paper, however, is on "if" conditionals.

Meaning²

There are many possible combinations of the clauses mentioned above, and the meanings they convey have subtle differences that can confuse even native speakers of English. The context in which a conditional is used has to be referred to in order to understand if you are dealing with a possibility, a regret, a wish, or an action that the speaker is willing to perform.
    
Celce-Murcia and Larsen-Freeman (1999) state that English conditional sentences express three different kinds of semantic relationships: (1) factual conditional relationships, (2) future (or predictive) conditional relationships, and (3) imaginative conditional relationships. Each of the three has its own sub-types.
    
Factual conditionals include four sub-types: (1) generic, (2) habitual, (3) implicit inference, and (4) explicit inference. Generic conditionals express relationships that are true and unchanging. They are normally expressed with the simple present tense in both clauses (e.g., "If you heat water to 100 degrees C, it boils") and are frequently found in scientific writing.
    
Habitual conditionals are similar to generic conditionals. Both types express relationships unbounded in time, but habitual conditionals are based on habit rather than physical law. Habituals express both present and past relationships. The same tense is used in both clauses. In most cases, it is possible in both generic and habitual conditionals to substitute "when" or "whenever" for "if."

By using "should," "happened to," or "should happen to," the negative quality of the "if" clause in hypothetical conditionals can be weakened more so the possibility of the result is stronger.

This weakening of the "if" clause does not occur in counterfactual conditionals because the "if" clause is strongly negated and the condition remains impossible.
    
Another difference between hypothetical and counterfactual conditionals has to do with time reference. Hypothetical conditionals can refer to the future as well as the present.

If Gandhi were alive, he would be shocked. (present)
If Gandhi had been alive during the 1960s, he probably would have been friends with  Martin Luther King. (past)

Celce-Murcia and Larsen-Freeman (1999) note that ESL textbooks and grammars often introduce the same three forms, which do not cover the complexity and variety of English conditional forms and meanings. The traditional "type 1, 2, and 3" conditionals noted by Celce-Murcia and Larsen-Freeman (1999:545) are:  

This is a problem area that can be utterly confusing to ESL/EFL students. Conditionals require coordination of verb forms in both the "if" and result clauses. This is complicated by the fact that verb forms in conditional sentences often do not retain their normal references to time. Gordon (1985:85)  states that "the past is used to represent improbability in the present or future, and the past perfect is used to refer to impossible events that didn't happen in the past." In addition, the present is used to refer to the future in the "if" clause of future predictive conditionals.
    
In order to explain the tense shifts in both clauses of conditionals, Celce-Murcia and Larsen-Freeman (1999) use The Bull Framework (see Bull, 1960), which posits four axes of orientation, or points of view, with respect to time: future, present, past, and hypothetical. Each of the first three axes in Bull's framework has a basic time slot in the middle and two possible marked slots -- one on the left signaling a time before the basic time of the axis and the other on the right signaling a time after the basic time of that axis. Bull's framework was originally designed for Spanish, but Celce-Murcia and Larsen-Freeman (1999:163) provide examples for English.

In terms of hypothetical conditionals, Celce-Murcia and Larsen-Freeman (1999:553) say

[W]e need two imaginative time lines for English so that we can account for both hypothetical present and future time and counterfactual present and past time. Also, we feel that the two imaginative axes are somewhat unique in that there is no "time before," simply a basic time and a time after (the perfect forms have other functions in these axes because hypotheticals and counterfactuals are shifted back in time to convey unreality). Thus we would represent the two imaginative axes as follows:

Number, Frequency, and Use of Conditionals

According to Hill (1960), there may be as many as 324 distinct tense-modal sequences in English conditional sentences. Several studies analyzing corpora of conditionals have been made, but Hwang's (1979) is frequently referred to. Hwang (1979) found that seven patterns accounted for about two-thirds of the conditional sentences in her spoken and written corpora. The frequency ranking for Hwang's (1979) corpora is shown below.

Fulcher's (1991) written corpus study of sentence forms using "if," most of which were conditional, identified 20 different "if" forms (see Appendix) in academic, narrative, and journalistic writing. The purpose of the study was to see if there were any differences in the frequency rate of "if" sentence forms once the variable of text-type was introduced; and if so, it might give some guidance regarding teaching conditionals depending on students' needs and purposes.
    
Fulcher's (1991:166) study found that the traditional types 1, 2, and 3 accounted for 61 (or 20.4%) of the 299 occurrences of "if" forms. The combination of if + present simple, present simple/continuous; if + present simple, imperative; and if + present simple, present modal totaled 135  (or 45.15%) of occurrences. The traditional type 3 conditional (if + past perfect, would have + -en) accounted for only nine (or 3%) of the total.
    
Fulcher (1991) also took the four largest categories and tested them for significance against text-types. Fulcher (1991:167) concludes that "there is indeed a link between student purpose and the need to learn particular conditional forms."
    
Fulcher (1991) suggests that teachers select which conditionals to teach depending on the most frequent forms in the type of text students are most likely to encounter. Students studying English for Academic Purposes (EAC), for example, might find that the traditional type 1 and 2 conditionals are less significant than other types. Fulcher (1991:167) believes that teaching context is also important to consider, as "the traditional 2nd conditional only seems to occur with any great frequency in narrative, which would presumably form a substantial part of a general English course," but not necessarily of other types of instruction.
    
Ford and Thompson (1986) claim that conditional sentences with initial "if" clauses, which account for nearly 80% of conditional sentences in their corpora, perform four functions in both oral and written discourse. The four functions are:

Order of Acquisition, Predicting Learners' Difficulties

Berent (1985) conducted two experiments comparing the production and comprehension of real, unreal, and past unreal (i.e., types 1, 2, and 3 referred to earlier) conditional sentences for 55 advanced and low-advanced adult ESL learners. The results of the experiment indicated that although past unreal conditionals were the most difficult to produce, they were the easiest for the learners to comprehend. On the other hand, real conditionals were the easiest to produce, but the most difficult to comprehend. Berent (1985:369) suggests that "form and function can be acquired at different times in the course of second language acquisition." Berent (1985) believes the results support the markedness theory of second language acquisition⁴.
    
Chou (2000), however, finds fault with Berent's (1985) assessment for production of the conditionals in the study. Berent's (1985) production test was a discrete-point, fill-in-the-blank test in which there was no meaningful context for each question. Participants filled in the verb form in the "if" clause or the result clause of a conditional according to the clue given by the verb in the other clause. Chou (2000) points out that the questions that contained hypothetical or present counterfactual conditionals can also be interpreted as past factual conditionals. Chou (2000) used two samples from the Berent (1985) study to illustrate his point:  

Chou (2000:66) also questions Berent's (1985) usage of the term "markedness" as being too vague:

[T]he term markedness is widely used in various contexts, such as typology, syntax, semantics, morphology, or discourse, and it has different meanings depending on the context in which it is used. However, Berent did not specify the context in which this term was used in his paper. It is possible he used the term markedness in the same way as the one used in linguistic typology research....

 Eckman's (1996) Markedness Differential Hypothesis (MDH) proposes that L2  difficulty can be predicted on the basis of the differences between learners' L1 and L2, and the markedness relationships that hold within those areas of difference....

Eckman also emphasized that "markedness refers to the relative frequency or generality of a given structure across the world's languages" (p. 198). He wrote, "Thus, markedness, in the sense used by the MDH, is not a matter of judgment or conjecture; it is an empirical matter" (p. 201).

However, I have not found any published linguistic typology research article that shows the markedness ranking of the syntactic representation of English   conditionals that are frequently taught in ESL or English as a foreign language (EFL) classes.

Maule (1988) believes the simplistic approach of introducing only type 1, 2, and 3 conditionals does more harm than good. If students are taught only a few simplified forms, they will be unequipped to encounter or express the variety of forms that exist in the English language. Their ability to process and produce the majority of acceptable conditional forms will be hindered.
    
Maule (1988) recorded a sample of 100 conditionals from U.K. television and found that the three traditional forms taught are not representative of the forms commonly in use. Only seven of his 100 samples conformed to the type 1 conditional. Maule's (1988:117) sample included 40 real non-past forms using "present tense, the imperative, modals, be to, have to, have got to, and going to," suggesting that learners' understanding of the type 1 conditional should be expanded to include other real non-past forms. Otherwise, learners may reject the other forms as unacceptable.
    
Maule (1988:121) takes issue with introducing students to the unreal past form (type 3) while ignoring the pattern for real past time conditionals (i.e., If + past, would), which students traditionally learn as the unreal present type:

It is a curious aspect of the traditional approach that while we are quite happy to teach students to deal with a situation which didn't happen [If it had snowed, we would have stayed at home], we somehow draw the line at equipping them with one which did [If it snowed, we stayed at home/would stay at home].... While teaching students that If + past simple, would + verb is used for unreal non-past conditions, we might also point out that it can be used for real past time conditions. If I had time, I would visit the museum could apply to either.

I do not teach the conditional clauses that form the majority of our samples, because as far as I can see they do not need much teaching: the tenses behave more or less as they should do: a present tense refers to present time, past to past, the imperative has its conventional significance -- and so on. Whereas in  our conventional "Three Conditionals", the tenses go haywire. The "First Conditional" is not just non-past: the present tense after the "If" actually refers   to the future, a totally anomalous use which my students find very difficult....

Similarly, the past in the "Second Conditional" refers to unreal present -- or at least non-past time, ... and the past perfect in the "Third Conditional" to the past....

In these constructions, my students, left to their own devices, come up with ungrammatical sentences, or sentences with meanings different from what they  intended.... They need to be taught the oddities of these particular kinds of conditionals; and if I simply expose them to the wide variety of conditional sentences available, ... I risk confusing them rather than helping them learn.

With all the problems associated with the difficulties of teaching and learning conditionals, as well as the call for the necessity of equipping students with more than the traditional three types used by many ESL textbooks and grammars, what should a classroom teacher do to teach these forms more effectively and efficiently? I would like to propose a system that, in order to make the connection between the actual time reference and backshifting, introduces the verbs "hope" and "wish" together with corresponding "if" sentence patterns and places conditionals into five convenient categories.
    
Celce-Murcia and Larsen-Freeman (1999) state that English makes use of several verbs of imagination such as "hope," "wish," "imagine," "pretend," and "suppose" that should be included in a comprehensive description of imaginative clauses. Of key interest to my proposal are the verbs "hope" and "wish." Celce-Murcia and Larsen-Freeman (1999:555) explain the connection between these verbs and conditionals:

The verb hope is similar to future (predictive) conditionals in that the same clauses that follow hope can also function as either the if clause or the result clause of a future conditional.... The verb wish, on the other hand, is similar to counterfactual conditionals in that the same clauses that follow wish can also function as the if clause or the result clause of a counterfactual conditional.

    
The proposal's "type 1 if" pattern combines Hwang's (1979) future predictive conditionals with her explicit inference factual or future with weakened result conditionals. This combination accounts for 24.6% of her written corpus and 19.9% of her spoken corpus. Various types (see Appendix, types 1, 6, 12, and 14) that comprised Fulcher's (1991) written corpus can be placed in the "type 1 if."
    
Hwang's (1979) present hypothetical or counterfactual conditionals combine with her present or future hypothetical or present counterfactual conditionals to make up the proposal's "type 2 if" category. These totaled 16% of her written corpus and 18.8% of her spoken corpus. Fulcher's traditional type 2 conditionals (see Appendix, type 2) made up 10.03% of his corpus.
    
The proposal's "type 3 if" category comprises from 3% of Hwang's (1979) to 4% of Fulcher's (1991) corpora and includes a "past cause, present result" pattern, which was not included in the above corpora. The "type 4 if" includes Hwang's (1979) generic factual conditionals, which accounted for 16.5% and 19.2% of her written and spoken corpora, respectively. Various types (see Appendix, types 4, 8, 10, 16, and 20) that accounted for 35.8% of Fulcher's (1991) corpus can be placed in the proposal's "type 4 if" category.
    
The first four categories of the proposed model account for 73.6% of Fulcher's (1991) corpus and 60.4% and 61.7% of Hwang's (1979) written and spoken corpora. The proposed model's "type 5 if" category can be used as a catch-all section to place all other "if" conditionals. The intent here is to raise students' consciousness concerning the wide variety of English conditional patterns that exist.
    
The proposed model for introducing "if" conditionals has several advantages. These include its relative simplicity, the consistency in the backshifting of tenses in both clauses in the first three categories, the coverage of the majority (from 60% to nearly 74% among the model's first four "if" types) of "if" conditionals found in Hwang's (1979) and Fulcher's (1991) corpora, the contextual connection with the verbs "hope" and "wish," and the raising of students' awareness that the traditional three types of conditionals do not cover the full range that the students' are likely to come across in their studies.
    
The main disadvantage would be the possible confusion for students in trying to interpret different semantic nuances within a single pattern. For example, the if + past, would pattern can be either a present counterfactional or a future hypothetical depending on the context. Instruction and practice in using context to determine semantic nuance are necessary.
    
The key points for the teacher to consider when introducing the patterns are (1) the "one step back" rule (i.e., backshifting) for verbs in the "hope" and "wish" sentences, for verbs in the "if" clauses in the corresponding "if" sentences, and for modals in the "result" clauses in the corresponding "if" sentences⁵; (2) the "hope" and "wish" sentence patterns are exactly the same as the "if" clauses in their corresponding "if" sentences; and (3) the importance of the time word at the end of each clause and sentence in order to focus students' attention on the time-frame and "one step back" rule.
    
Clearly, students' levels should be taken into consideration in determining whether to introduce the entire model at the same time or only portions of it. Celce-Murcia and Larsen-Freeman (1999) suggest that acquisition of the English tense-aspect system, the modal auxiliaries, and negation is a prerequisite for ESL/EFL students to acquire the full range of English conditionals.
    
The proposed model is not the perfect answer to the conundrum of how to teach conditionals, but for the daily practical needs of teachers and students in the classroom who are under time constraints it provides a practical starting point that goes beyond the traditional approach of introducing only three types. A forthcoming paper will focus on detailed lesson plans based on the model.

1. Comrie (1986: 94) uses "backshifting" to mean the "use of a morphologically past tense with present (or future) time reference and of pluperfect with past time reference." Comrie (1986) notes that backshifting is a common phenomenon in Indo-European and European-area languages in conditionals with a high degree of hypotheticality, such as counterfactual conditionals.

2. The information in this section is taken mainly from Celce-Murcia and Larsen-Freeman (1999).

3. This table is adapted from Celce-Murcia and Larsen-Freeman (1999:557).

4. Eckman's (1977) original Markedness Differential Hypothesis (also known as "markedness theory") accounted for relative degrees of difficulty by means of principles of universal grammar. Celce-Murcia and Hawkins (1985:66) summarize markedness theory as follows:

It distinguishes members of a pair of related forms or structures by assuming that the marked member of a pair contains at least one more feature than the unmarked one. In addition, the unmarked (or neutral) member of the pair is the one with a wider range of distribution than the marked one. For example, in the case of the English indefinite articles (a and an), an is the more complex or marked form (it has an additional sound) and a is the unmarked form with the wider distribution.

5. There is a consistent progression backward from present to past to past-past (i.e., past perfect) in the "if" clauses (and "hope/wish" sentences) and "will/can" to "would/could" to "would have/could have + -en" in the result clauses of the first three types in the proposal.

I would like to express my heartfelt thanks to my colleague Dominic Marini for his invaluable proofreading of this paper.

Berent, G. P. 1985. "Markedness Considerations in the Acquisition of Conditional Sentences." Language Learning, 35/3:337-372.

Bull, W.  1960.  Time, Tense, and the Verb: A Study in Theoretical and Applied Linguistics, with Particular Attention to Spanish. Berkeley: University of California Press.

Celce-Murcia, M. and D. Larsen-Freeman. 1999. The Grammar Book: An ESL/EFL Teacher's Course (2nd Edition). New York: Heinle and Heinle.

Chou, Chen-Ling. 2000. "Chinese Speakers' Acquisition of English Conditionals: Acquisition Order and L1 Transfer Effects." Second Language Studies, 19/1:57-98.

Comrie, B. 1986. "Conditionals: A Typology." In Traugott, E. C., Meulen, A., Reilly, J. S., and C. A. Ferguson (Eds.), On Conditionals. Cambridge: Cambridge University Press.

Covitt, R. I. 1976.  "Some Problematic Grammar Areas for ESL Teachers." Unpublished M.A. thesis in TESL, UCLA.

Eckman, F. R. 1977.  "Markedness and the Contrastive Analysis Hypothesis." Language Learning, 27:315-330.

Eckman, F. R. 1981. "On the Naturalness of Interlanguage Phonological Rules." Language Learning, 31:195-216.

Eckman, F. R. 1996. "A Functional-Typological Approach to Second Language Acquisition Theory." In Ritchie, W. C. and J. K. Bhatia (Eds.), Handbook of Second Language Acquisition. San Diego: Academic Press.

Ford, C. E. 1997. "Speaking Conditionally: Some Contexts of If-Clauses in Conversation." In Athanasiadou, A. and R. Dirven (Eds.), On Conditionals Again. Amsterdam: John Benjamins.

Ford, C. E. and S. A. Thompson. 1986.  "Conditionals in Discourse: A Text-Based Study from English." In E. Trangott, et al. (Eds.), On Conditionals. Cambridge: Cambridge University Press.

Fulcher, G.  1991.  "Conditionals Revisited."  ELT Journal, Vol. 45/2:164-168.

Gordon, D. 1985. "The Marking of Conditionality in One Learner's Interlanguage: A Semantic Analysis." Working Papers, Department of English as a Second Language, University of Hawaii, Manoa, 4/1:77-102.

Halliday, M. and R. Hasan. 1976. Cohesion in English. London: Longman.

Hill, L. A. 1960. "The Sequence of Tenses with If Clauses." Language Learning, 10/3 and 4:165-178.

Hwang, M. 1979. "A Semantic and Syntactic Analysis of 'If' Conditionals." Unpublished M.A. thesis in TESL, UCLA.

Maule, D. 1988. "'Sorry, But If He Comes, I Go': Teaching Conditionals." ELT Journal, 42/2:117-123.

Mayes, P. 1994. "Conditionals and the Logic of Desirability: An Interview with Noriko Akatsuka." Issues in Applied Linguistics, 5/2:449-461.

Quirk, R. and S. Greenbaum. 1973. A Concise Grammar of Contemporary English. New York: Harcourt Brace Jovanovich.

Ur, P. 1989. Letter in Response to Maule. ELT Journal, 43/1:73-74.