TESOL News Vol. 8 No. 3 1987
EVALUATING LINGUISTIC DIFFICULTY
Thor May 1987 University of Technology Lae, Papua New Guinea
Note 1: This paper was first published in TESOL News Vol. 8 NO 3 1987 (Australian Commonwealth Schools Commission). The analysis it contains is still very relevant to ESL teachers.
Note 2: At the time of writing Thor May, was preparing a doctoral dissertation in theoretical linguistics, and teaching in the Department of Language and Communication Studies at the Papua New Guinea University of Technology. Since 1976 he has taught English to speakers of other languages in secondary, adult and tertiary fields in Australia, New Zealand and Papua New Guinea, as well as China and South Korea. He has lectured in formal linguistics at the University of Newcastle (NSW), Southern Cross University (NSW), The University of the South Pacific (Suva, Fiji). and Pusan University of Foreign Studies (Busan, South Korea). Thor May has a doctorate in language teaching productivity from the University of Newcastle, NSW.
Abstract: While ESL teachers cannot eliminate linguistic difficulties, with an awareness of the factors involved it is possible to minimise the confusion of their students. This article systematically analyses some important problem areas in language learning. It itemizes a range of syntactic and semantic phenomena, considering in each case how the rule or pattern might pose a difficulty for some learners. This paper has been published for a number of years now, and the writer has become aware that many teachers themselves have found it a useful aid in preparing and presenting course material.
Table of Contents
INTRODUCTION // orders of complexity // LEXICAL DIFFICULTY // Syllabic length:// Clusters // Irregular spelling // Irregular stress // Affixes // Multiple denotation // Range of connotation // Specialized application // Frequency of lexical items // Selectional restrictions // Subcategorical restrictions // MEASURES OF STRUCTURAL COMPLEXITY IN SENTENCES // Sentence length // Qualifying words // Adverbial and prepositional phrases // Conjunctive sentences // Equi-deletion // Deletion by convention // Permutation // Transposition // Embedding // Sentential complements // Topicalization // Presupposition // Tense // Aspect // Agreement (concord) rules // Anaphoric, cataphoric and exophoric references // DISCOURSE COHESION // CUEING // IDIOM // CONCEPTUAL DIFFICULTY // More accessible reference // Less accessible reference // Types of Inference // REFERENCES
Beginning language teachers learn things they never expected to learn, sometimes to their astonishment, from the confusion of people accustomed to incomprehensible jargons like Serbian or Vietnamese. These teachers discover that their own, familiar, user-friendly English is also a constant source of bafflement. Worse, they find it extremely difficult to explain just what is going on in the language. The trick is to put one's finger on just what it is that so puzzles non-native speakers. Consequently, there is nothing as troublesome for an ESL teacher as guessing what written material can be safely used, what language can be uttered, and what kind of response can reasonably be expected in return.
The purpose of this article is to identify some of the things that make language difficult. If you are a teacher, your students will not usually be able to express these problems, and there is only so much that you can do to adjust. Being aware of what is involved though can make things a bit less haphazard for everybody.
From time to time attempts have been made to set boundary markers for measuring difficulty in natural languages: For example, various formulae are available for determining the "reading age" of written material. These are used by reading teachers, as well as the editors of newspapers. Afternoon tabloids supposedly have a reading age of about eleven years. The formula might give a number to x decimal places, but humans being what they are, such measurements are very rough and ready indeed. The reason for their unreliability can be more formally described: unlike the context-free, closed languages of, say, mathematics and formal logic, the systems of natural language are so complex and adaptable that reliable methods of characterising their interaction, let alone assigning weighted judgement to "difficulty" for users and learners, are not yet within reach.
There is no reasonable equation that we can apply to any group of complex language factors, nothing that permits us to say "X + Y+ Z means that this sentence is too difficult for these users". Such evaluation remains more in the realm of an art than a science: it depends upon sensitivity, the direct contact with particular speakers which text-book writers don't have, and experience. I can offer one encouraging comment from my own observation: where the intrinsic content of material is of great personal interest to the end-user, the level of structural complexity can often be escalated without destroying the message.
In what follows I have analysed some of the many factors which complicate the interpretation of language for all people, but particularly those working in a foreign tongue (such as English for immigrants). Please note that a) this list is in no sense exhaustive; and, b) it is not compiled as a check-list of teaching points.
Think of the analysis below as a mental starter kit. What is difficult for particular students will depend upon their stage of English acquisition at a given time. The kind of fascinating work being done by Pienemann and Johnston ( 1986) through Adult Migrant Education Services (AMES) in Sydney may eventually make it easier for us to be clearer about the acquisition stages learners are at.
Briefly, these researchers propose that language acquisition processes divide into developmental and variational features. Developmental features are acquired in a particular sequence while variational features are much more personal, depending upon things like attitude to the host culture. Taken together, these two types of features mean that there are many possible paths to successful language acquisition. Nevertheless, the developmental element has allowed Pienemann and Johnston (after Clahsen, 1984) to suggest an order of complexity for the mental processing of language. It is as follows:
1. Single words only.
2. W X Y Z canonical order.
3. (W) X Y (Z) initialization/finalisation. Final elements can be moved into initial position or vice versa; that is, from salient position to salient position.
4. W X (Y) Z semi-internal permutation. Internal elements are permuted, but only into a salient (initial or final) position.
5. W (X) (Y) Z fully internal permutation. Salient positions are no longer used as a means of orientation for permuted elements.
Although I suspect that the whole truth may be rather more subtle than 1 to 5 above, the general approach is very promising. Pienemann and Johnston are looking for ways to use this "complexity list" as a tool for assessing the L2 development of AMES students. For reasons that I won't go into, they argue that it has great value where measures such as the ASLPR (Australian Second Language Proficiency Rating) fall down.
What does all this have to do with our purpose here? Well, if language complexity becomes a widely applied tool for assessment and prediction in Australian ESL classrooms, then teachers will have to come to a more organised understanding of what actually makes language difficult. Read on!
B. An Analysis Of Linguistic Difficulty
1. LEXICAL DIFFICULTY
(a) Syllabic length: Beginners in a foreign language rely heavily upon short-term memory, and often find long words (and long sentences) difficult to process.
(b) Clusters: Consonant clusters (e.g., in a word like "script") are fairly common in English but rare in many languages. The rules for their use in English words are complicated and they are difficult for others to pronounce (just as we find Polish clusters difficult).
(c) Irregular spelling: The link between sound and written symbol in English is erratic at the best of times. The symbol "e" for example has about thirteen pronunciations in English. L2 learners become very demoralised by this kind of insanity!
(d) Irregular stress: English word stress does have a degree of regularity. The penultimate syllable (2nd last) is stressed in many words, the ante-penultimate (3rd last) in many others. And there are many exceptions; (check out this paragraph!). English sentence stress can be varied in remarkable ways. There are many languages in which life is less complicated than this.
(e) Affixes: English makes use of many prefixes and suffixes, mostly to influence meaning rather than syntax. Some affixes are highly productive (e.g. un in "unhappy"), others are much rarer (e.g. a in "atypical").
(f) Multiple denotation: Any competent English speaker can think of several meanings for, say, "make" or "pick up". With L2 learners it is always a good idea to ask, "Is this particular meaning of the word new for these people; is it rare?"
(g) Range of connotation: All words (and sentences) have both explicit and implicit meaning. The implicit meaning usually depends upon shared cultural knowledge or experiences. The connotations of "tree", "fish" and "school" are quite different for Australians and, say, Solomon Islanders. Now whose set of connotations did you have in mind when you last used those words?
(h) Specialized application: Every trade, science and interest group applies specialized. meaning to common words, (e.g., this analysis!) Scan your teaching material for specialized usages that you have taken for granted.
(i) Frequency of lexical items: Rare words need not be excluded from L2 material if they are also important. But be selective. A good language student under ideal conditions can "absorb" several new words on a given day. However, getting the learner to recall and use the "several new words" encountered in each of the preceding thirty - or three hundred - days is asking a lot.
(j) Selectional restrictions: "Colourless green ideas sleep furiously" sounds rather odd. "Idea" is a word not normally associated with "green" or "sleep" in English culture (although poetry often employs such juxtapositions). Such selectional restrictions vary greatly among languages. Be sensitive to the unexpected puzzlement of your users over "obvious" things.
(k) Subcategorical restrictions: Every word is sensitive to its linguistic environment. Rigid word order rules in English create quite categorical restrictions on words. Thus only certain classes or words can occur to the immediate left of, say, "eat" : usually it is an animate noun. Some words, especially archaic ones, have highly restrictive environments. For example the rare verb, "slough"; almost always occurs with the preposition "off". (All verbs, incidentally, are associated with limited sets of prepositions). With L2 users, try to avoid words with very restrictive environments, or at least explain their limitations.
2. MEASURES OF STRUCTURAL COMPLEXITY IN SENTENCES
(a) Sentence length: Short sentences are much easier for non-fluent as well as less able L2 users (also L1 users!) to process. This applies to both the spoken and the written medium. Be aware, however, that spoken (as opposed to written) language offers no opportunity for user-review when extracting meaning.
(b) Qualifying words: Nouns in English may be qualified by determiners (the, a), adjectives (old, big), quantifiers (many, few), intensifiers (very) ... etc. Verbs may also be qualified in various ways. A large number of qualifying words can confuse and distract insecure L2 users.
(c) Adverbial and prepositional phrases: Sentences may be qualified by various phrases and there is often a degree of freedom about their placement in the sentence:
With a knife Fred cut the salami. Fred, with a knife, cut the salami.
Fred cut the salami with a knife.
The multiple use of adverbial or prepositional phrases, or their unusual placement, can confuse L2 users. On the whole, they are best put at the end of sentences. Not always though: "On the whole" in the last sentence can only occur at the end with special pausing and stress placement.
(d) Conjunctive sentences: And, but, and or are the commonest conjunctions used for joining English sentences, and their application may often (not always) correspond to the functions of logical operators. L2 users have least trouble with and, a little more with but and or. Writing L2 material, it is often possible to reduce a long sentence with conjunctions to several shorter sentences. This lightens the processing load. The second sentence in a conjunctive set often has confusing deletions (see "Equi-deletion" below).
(e) Equi-deletion: Elements in a conjoined sentence are often deleted in English. The grammar may even require that they be deleted:
1. The bomb demolished the car and [the bomb demolished] the shop window.
2. Fred can come if he likes and so can Mary [come if she likes].
Any listener must "recover" these deleted elements in order to interpret the sentence. This may be beyond the competence of insecure L2 users.
(f) Deletion by convention: A great deal is left out of everyday language on the basis of shared understanding. This is an absolute minefield for L2 users, since the rules for omission (where they exist) are often obscure:
+ How would you like to come and get drunk with me?
- If you're shouting [the drinks], I wouldn't mind
[coming and getting drunk] at all.
(g) Permutation: The most frequent way of permuting subject and object noun phrases relative to the verb in English is the passive "transformation":
NP1 V NP2
Fred bit the dog
NP2 BE V en BY NP1
The dog was bitten (by Fred).
Children, certain kinds of aphasic adults, and L2 learners can all experience difficulty decoding passive sentences, often assigning an active meaning. There is a good deal of individual variation in this decoding "blockage", and some passives, by virtue of their meaning, are much less likely to be misinterpreted than others. It is wise, however, to cast a very careful eye over curriculum material.
(h) Transposition: It is possible for whole clauses to be fronted in English:
1 a) [For the rain to come now] would be a disaster. . . is a transposed form of ....
l b) It would be a disaster for the rain to come now.
2a) [That orchids grow on Guadalcanal] is well known. ... is a transposed form of ....
2b) It is well known that orchids grow on Guadalcanal.
The interpretation of such structures requires that users see the whole transposed clause as subject of the main verb (i.e. of BE in this case). This imposes a heavy processing burden. Clefts (the it form in the b) sentences) have their own problems of course (see "Clefts" below).
(i) Embedding: Written English, especially, makes extensive use of sentences embedded or "nested" within other sentences. The normal syntactic device employed is the relative clause. There is a limit to the capacity, even of native speakers, to process embedded clauses/sentences. Does the following give you vertigo?
He knew [who the man] [who was wearing a yellow raincoat] [which came from the Woolworths][which is in Magellan Street] was.
The ability of L2 speakers to process even single embeddings before they are fairly fluent is quite limited. In spoken language such structures can create extreme difficulty.
(j) Sentential complements: These are complex sentences of the kind:
1. [I believe [that the sky is pink tonight]] ... where that is the complementiser.
These forms really embed a second sentence as a complement of the matrix verb (i.e. of believe in this case.) The word that in such sentences is called a complementiser. There are at least three kinds of complementisers in English, (THAT, FOR-TO, POSS'-ING) but not all three can go with every matrix verb:
2. [I believe [elephants to have large appetite]] ... where to is the complementiser.
3. *[I believe [Wendy's playing of the flute]] .. where 's + ing is the complementiser.
4. *I like [that the sky is pink tonight]
I like [elephants to have large appetites]
I like [Wendy's playing of the flute]
Thus the combination of complementisers available to each matrix verb appears to be quite idiosyncratic. (Various linguistic models try to identify reasoned regularities, but they are hotly disputed by professional linguists themselves.)
Complex sentences employing matrix verbs are of course far more difficult than simple sentences for L2 users to decode. Conscious rules for their use are rarely taught in language courses (perhaps because the "rules" are so little known) and they can pose enormous difficulties for L2 speakers trying to generate standard language. Nevertheless matrix verbs cannot be entirely avoided since some of them have a very high frequency in normal English usage.
(k) Topicalization: The simplest and commonest of English sentences have the structure:
Noun phrase [Subject] + Verb + Noun Phrase [Object]
The Subject is typically the concept or thing about which the Object provides new information. Sometimes, however, it is desirable to emphasise this information focus. Hence the colloquial:
Bill, he is a disaster! ..where Bill is the topic focus
( Bill and he refer to the same person).
Sometimes it is necessary to shift the focus altogether. Passives achieve this:
1. Fred bit the dog
... where Fred is the topic-focus, and the dog is the information-focus
2. The dog was bitten by Fred.
... where the dog is the topic-focus, and Fred is the information-focus
Note that unusual stress placement can alter these interpretations.
Clefts also achieve an unusual topic focus:
[It would be a disaster] for the rain to come now.
... where it is a neutralizing cleft, and a disaster is the topic-focus.
So-called pseudo-clefts also have this effect:
[What we need] is immediate action
... where what is a neutralizing pseudo-cleft, and need is the topic-focus.
The perceptive reader will think of other topicalizing techniques in English. The important point for our purposes here is that they all represent a shift from the standard patterning and stress placement of simple Subject + Verb + Object; that is, the. English sentence patterns most familiar to L2 users. They may therefore slow down or confuse decoding of the message.
(1) Presupposition: Presupposition plays a crucial part in the interpretation of natural language. Sometimes it is signalled unequivocally by the matrix verb:
la. I realised that you disliked him.
lb. I didn't realise that you disliked him.
Both a) and b) presuppose that "you disliked him". Sometimes the presupposition is less certain, although likely:
2. I won't be here after 5pm.
... probably (but not necessarily) presupposes that "I will be here before 5pm". It is asking a lot to expect people struggling with a second language to draw reliable inferences of this kind, for both linguistic and cultural reasons, and material should be surveyed for such unintended complications where they are important to the meaning.
(m) Tense: Formally speaking, English has only two verb tenses: NON-PAST (unmarked) and PAST (marked with the suffix -ed on regular verbs). The future is indicated by a modal auxiliary like WILL. The most frequently used verbs, unfortunately, have irregular PAST forms. Nevertheless, compared to many languages, the inflections of English are relatively simple. Decoding English tenses is not likely to be a major problem. There will be the usual learner errors in creating new sentences, particularly where tense agreement is necessary in compound sentences with more than one verb.
(n) Aspect: If English has a relatively simple tense structure, its use of ASPECT is often subtle and baffling. ASPECT indicates the relationships in time between events or items in the sentence, sometimes with the speaker as a reference point.
1. Continuous Aspect
a) He is cleaning his teeth.
This is the most typical use of the continuous aspect. It describes a process or activity, normally of limited duration.
b) I'm going.
This indicates future intention in the sense of I am about to go.
c) The postman is calling at number twenty-six again.
This use implies a repeated action. Note that it can differ subtly from the English simple present tense, which normally describes a habit:
i) The postman calls at number twenty-six (every day).
ii) ?* The postman calls at number twenty-six again.
(Just to confuse things, the form of simple present shown in sentence ii) is sometimes used in commentaries, stage directions, etc. where a continuous aspect would normally be applied. Listen carefully to a few sports broadcasts.)
2. Present Perfective Aspect
He has eaten lunch
... describes an event that occurred some time before NOW. It can't usually describe a single exact point in time:
*He has eaten lunch at 12 noon today.
3. Past Perfective Aspect
i) He had drunk three cups of coffee by 1 pm.
... describes an event that occurred at some time before a past point of time. It can't (usually) describe the exact time of occurrence.
ii) ? He had drunk 3 cups of coffee at 1 pm.
(There are exceptions to the rule, however, just to confuse things. It is possible to give sentence ii) an interpretation).
There is no way that any realistic use of language can avoid fairly extensive use of tense and aspect. It must be expected however that L2 users will often come adrift on the subtleties of aspect, so teachers should at least arm themselves with some explanations.
(o) Agreement (concord) rules: The possibility of confusing tense agreement in compound sentences has already been mentioned. Gender is largely eliminated in modern English, except in the pronominal system (I, he, she, it, you, we, they ... etc.), so serious difficulty is unlikely. Some problems do arise with number agreement in mass and count nouns (some milk; several apples). These are learning problems of production (encoding) rather than sources of confusion in decoding.
p) Anaphoric, cataphoric and exophoric references: These are the devices in a language which permit reference to something previously mentioned in the text (or dialogue), something yet to be specified, or something external to the text altogether.
1. Anaphoric items refer to a previous reference:
e.g. Wash six apples. Put THEM in a flat dish. ( six apples <= them)
2. Cataphoric items refer to something yet to be specified:
e.g. If HE's not careful, John will lose his job. (he => John)
3. Exophoric reference refers out of the text altogether:
e.g. IT looks rather bare. (Context: The new office ...)
Not only pronouns are anaphoric. Demonstratives (this, that, these, etc.), definite articles (the), comparatives (more...than...etc.) all have an anaphoric functions. Likewise a whole category of sequencing words like: therefore, however, nevertheless, so that, then, after that ... and so on.
The effect of anaphora in discourse is to create cohesion. Beyond sentence-length, they are the only language markers showing cohesion, holding the dialogue together.
Of course, the anaphoric devices of each language are different, and they are rarely taught to the ESL learner. He is thus likely to miss many of the cues available to a native speaker, and be unaware of how to generate them in his own speech. A special problem for the learner is multiple anaphora - for example, the repeated use of a pronoun like it, sometimes with different references in the same passage.
3. DISCOURSE COHESION
The basic unit of linguistic organisation in all languages is normally taken to be the sentence. However, sentences themselves generally occur in longer strings of utterance or text, and their placement in these larger entities is not random.
The role of anaphora in discourse cohesion has already been mentioned. While anaphora disambiguate references and sequence in a text, anaphor density also imposes a processing load on the user, whose mind has to recall or anticipate other elements of language. For the language learner, struggling to hold half a dozen words of the new language in short-term memory, an utterance with high anaphor density may remain partly uninterpreted. Similarly, ambiguity in anaphor reference may well defeat him. One of the more certain measures of fluency in a language is the facility with which anaphora are manipulated and decoded.
The sheer length of utterance, or paragraph length, are important considerations in making language accessible to an ESL learner. In general, the briefer the text, the more accessible it is to the insecure user, again for reasons of processing/memory.
Internal organisation in a text may of course be more, or less, opaque. There are many ways of developing an argument; for example, deductive, inductive, thematic, empirical or eclectic approaches. Each has its own conventions, and many of these conventions are culture-specific. Thus the English style of deductive argument is largely unknown in many societies. It is as well for those who hope to teach or evaluate ESL learners to be familiar with prevailing conventions in the source language, and to be quite explicit about their own systems.
Cueing in language is the process of signalling to a user what is about to be said. It is part of a much more general process. There are essentially two time frames within which an organism can cope with any problem:
(a) Response on impact frame : the event has not been anticipated. It is therefore handled reflexively, or held in memory while a response is being formulated; or
(b) Event anticipation frame: alert itself to the likely important features of that event, and hypothesise an adaptive response. This second style of problem solving is by far the more efficient, and is characteristic of higher-order animals which normally seek to maximise this style of experience.
Language use is a supreme example of anticipatory behaviour. Most of the formal devices of syntax unconsciously cue a competent user to the structures which will follow. Anaphora is wholly concerned with this aspect of language. Similarly argument, attitude development and other kinds of meanings are typically anticipated by a listener from instant to instant.
Needless to say, the learner of a language is severely disadvantaged in all kinds of anticipatory interpretation. Not only is he unfamiliar with the formal conventions of the language, but the culture and shared knowledge on which its use is based are probably unfamiliar as well. This adds up to a problem of surpassing complexity. To be effective, language material designed for ESL learners must incorporate cues for the decoder which are far more explicit than those needed with a native speaker. Some common cueing techniques in ESL teaching are:
- previous use of the material.
- answer cues embedded in questions.
- situational cues (physical objects).
- other contextual cues (e.g. winking, shrugging, leaving the room).
"It is my earnest desire that you enter into conjugal relations with my person" would be an odd way to propose marriage in English. Grammaticality is a necessary but not sufficient condition for felicitous language use. For every L2 speaker, English is a minefield of idiomatic usage: things that seem logical can't easily be said (e.g. the marriage proposal above) whereas other "normal" expressions seem insane. For example, "I'll tell on you" can be quite incomprehensible to L2 learners unless clarified by context or by explanation.
Hence L2 teachers must always attempt to see language (including their own statements) through the foreign eyes of students. There is value in learning idioms, but their density in any given utterance must be controlled.
Idiom in language is not a simple or clearly defined concept. Idiom really means that some group of speakers accept some collocation of words as a very familiar way to express a certain idea. Sometimes that collocation may become a fixed formula unrelated to the original meaning of the individual words (e.g. "kick the bucket" = to die). More often however the fixedness shifts both over time and for different groups of speakers. There is a huge range of phrases which we don't really define as idioms, but which to a greater or lesser extent are familiar to the daily users of any language. From the learner point of view, this broader sense of discovering how particular ideas are usually said is more important than learning a handful of exotic "idioms".
6. CONCEPTUAL DIFFICULTY
Many English-language students are apt to find themselves dealing with concepts more difficult and more advanced than those normally encountered in their own cultures. This is especially the case for students from small or emerging nations who are forced by lack of indigenous books and facilities to pursue their career training in an English language medium.
Conceptual difficulties play a large part in setting the learning horizon for native speakers too, of course. But often a difficult idea can be disseminated to a wide audience by the judicious use of language. Two of the controlling factors especially pertinent to L2 teachers in this kind of language management are ease of reference and type of inference.
a) Factors affecting ease of reference:
More accessible reference
Less accessible reference
- class/generic type
b) Types of Inference
- logically necessary/formally presupposed (e.g. a syllogism)
- culture specific (e.g. "He beckoned" --> , therefore He wanted X to approach)
- specialised to topic/discipline
(e.g. "The NP was immediately dominated by an S' ,... therefore It was a subject NP. )
- explicit in the utterance
(e.g. "I avoided the intersection" -- because it was dangerous)
- implicit in the utterance
(e.g. "She had a black eye" --- therefore. something/someone had hit her)
- personally verifiable (e.g. "We're getting wet"--because it is raining.)
- verifiable in principle
(e.g., "The water-table is at a record low" ....because There has been a drought.)
This list of inference types is illustrative rather than exhaustive. The alert teacher will see that each imposes its own special demands upon the L2 user. Many inferential processes are extended, some with hidden presuppositions, and the issue then becomes how many steps an insecure L2 user can process without losing track of the anaphora and the argument.
H. Clahsen, ( 1984) "The Acquisition of German Word Order: A Test Case from Cognitive Approaches to L Development" in Anderson, R. W (ed.) ( 1984) Second Language: A Cross Linguistic Perspective; pub. Rowley, Mass., Newbury House; 2I9-244.
H. Clahsen, (1985) "Profiling Second Language Acquisition": in Hyltenstam, K. & M. Pienemann eds.) Modelling and Assessing Language Development. pub. Clevedon, Avon, Multilingual Matters; 283-33I.
M. Pienemann, & M. Johnston, ( 1986) "An Acquisition Based Procedure For Language Assessment" in Australian Journal of Applied Linguistics, Vol.9, No.l, 1986, pp 92-122.
Evaluating Linguistic Difficulty (c) Thor May 1987, 2012; all rights reserved To e-mail Thor May, thormay AT yahoo.com *** Links *** *** Home Page *** go to Technical Stuff*** [top of page] end