Lost but not found, or: when grammarians fail us: The case of (past) participial premodifiers in English

As a teacher of English to speakers of German, I am routinely confronted with mistakes involving the misuse of participial premodifiers. Like many other languages possessing participial forms, German places only minimal constraints on their prenominal use. The constraints in English, however, are both intrusive and perplexing. English happily allows 'lost property' or 'broken promises' or 'hired car' but normally disallows *'found property' or *'kept promises' or *'bought car'. Unaware of these constraints, my students commonly overuse participial premodifiers in English, producing errors that are easy to recognize and correct, but hard to elucidate. Focussing on the prenominal use of past participles, the present article tries to show that this is one area where language teachers are being seriously let down by grammarians. The standard pedagogical grammars provide scant guidance and even reputable scientific grammars disappoint. Only one leading grammar attempts to address the problem, but the account it provides is fragmentary and largely unconvincing. Latching on to the one aspect of this account that is correct, I try to develop a unified account of the phenomenon that may serve as a basis for a more satisfactory treatment of the issue both on the scientific and the didactic level.


Introduction
English is without doubt not only the most widely studied but also the most extensively researched and best documented language in the world. Teachers of English as a foreign language have at their disposal an unparalleled array of resource materials -textbooks (from general to highly specialised), dictionaries (monolingual and bilingual), grammars (single-language and comparative) -many of which have helped to set the international standard in their field. It must therefore seem churlish for an English teacher like myself to suggest that grammarians might be failing us. While it is probably true to say that no grammar will ever be complete or based on a definitive explanatory model, it seems almost inconceivable that what chinks there may still be in the linguistic description of English should be large enough to embarrass the classroom teacher. Surely all the grammatical structures we need to teach even at the most advanced level have been comprehensively described and made accessible in a wealth of didactic forms.
However, anyone who has perused R. M. W. Dixon's recent critique of English dictionaries will feel less confident (Dixon, 2018). In this devastating study, Dixon, one of the world's foremost scientific linguists 1 , argues persuasively that the entire English tradition of lexicography, from Robert Cowdrey and Samuel Johnson to Noel Webster and the OED, is fundamentally flawed. If a dictionary's basic task is to tell you when to use one word rather than another, then English dictionaries are failing us, because this is precisely the information that is systematically lacking. According to Dixon, lexicographers have fallen into three traps: regarding words as self-contained entities; paying insufficient attention to grammar; and relying too heavily on definitions ('glosses'). We cannot take issue with this claim here, but we should at least take heed. If it is conceivable that English lexicographers are failing us, and doing so in a way that directly affects ordinary dictionary users, then it is surely not inconceivable that English grammarians might be failing us too, and in a way that directly affects ordinary teachers and learners.
The specific grammatical issue that I would like to focus on is participial premodification, i.e. the use participles in a prenominal position. To make the topic more manageable, I shall restrict myself to past participles, because these are the ones that seem to cause most trouble. I make no apology for the fact that I intend to approach this issue as a teacher, from a teacher's perspective, and not as a scientific linguist, which I am not 2 . While I have conducted an extensive search for relevant literature, consulting dozens of pedagogical and scientific grammars (monolingual and bilingual) and searching numerous databases, I do not claim to have researched the issue as thoroughly as I would have had to, were I a professional grammarian. For any classroom practitioner, there are obvious limits to the time and effort that can be devoted to such a quest.
The students I have taught for more than twenty years are young engineers for whom German is usually either their mother tongue or their principal language of daily use. However, what I have to say applies equally to any academic specialism and to a variety of linguistic backgrounds. But it is likely that speakers of Germanic and Slavonic languages will be affected most since these languages allow the use of participial premodifiers while placing only minimal constraints on their use.
In German, the past participle of any verb can be used in a prenominal position as long as the result makes plausible sense. Zero-valency verbs (schneien, regnen) are systematically excluded (*geschneite Landschaft, *geregneter Tag) but this constraint is intuitively obvious to native and non-native speakers alike because the result is patently nonsensical (in such contexts German uses monovalent verschneien, verregnen instead). In English, the constraints on the prenominal use of past participles are both intrusive and perplexing. We talk about lost property but not *found property, broken promises but not *kept promises, deposited money but not *withdrawn money, about incurred losses but not *gained profits, required conditions but not *needed conditions. The constraints operating in these cases seem perfectly natural to native speakers but are often obscure to non-native ones, particularly in the absence of any formal instruction on the matter. Not having been alerted to the issue by teachers themselves often wholly unaware of it, students respond by over-generalizing participial premodification, especially where their native language routinely allows, even encourages, such premodification, while placing only minimal restrictions on it.
In my own work, I regularly ask my students to write short texts relating to their engineering project work, and when they do, I am resigned to receiving dozens of phrases like this: *the defined problem *the bought equipment *the arisen difficulties *the overcome obstacles *the conducted experiment *the obtained results Dyirbal (Pama-Nyungan, N. Queensland) is commonly regarded as setting a benchmark in the description of hitherto unrecorded languages.

*the found solutions
Of course, I dutifully correct them, usually by converting the prenominal participle into a postnominal relative clause (e.g. the problem we defined, the difficulties that arose), but find myself hard put to explain my intervention. I feel like saying, 'Don't use a past participle before the noun like that', without being able to say what I mean by like that. But like that is of the essence, because it is clearly not the case that English has any special aversion to participial premodification. In fact, we seem to like participial premodifiers so much that we can hardly imagine breakfast without them: fried bacon poached eggs boiled tomatoes grilled sausages baked beans steamed mushrooms smoked herring Even if we manage to get round the 'smoked herring' by calling them 'kippers', the difficulty remains: why are participial premodifiers fine in the second case but not the first? Or to put it another way: what (grammatically speaking) do poached eggs and grilled sausages have that *overcome obstacles and *obtained results do not?

Discussion
What is immediately obvious is that the answer is not obvious, and the further you look, the less obvious it becomes. Baffled, you may decide to consult whatever pedagogical grammars are to hand. This experience is instructive but hardly in the way intended. The standard pedagogical grammars (Murphy, 2012;Hewing, 1985;Swan, 2009;Emmerson, 2010 etc.) rarely acknowledge the issue. Swan simply tells us that the participle can be used before the noun and after the noun (Swan, 2009: 380-1). The only one that makes any serious attempt to address the matter is Hewing. This is what he says (Hewing 1985: 170) (I omit references to present participles, which are not my immediate concern here): A few participles are used immediately after nouns but rarely before them […]: […] • My watch was among the things taken (but not … the taken things) Other participles like this include caused, found, provided, used.
Some participles can be used before or immediately after nouns. For examples we can say: • Rub the area infected with this antiseptic cream. or • Rub the infected area with this antiseptic cream. […] Other participles like this include affected, broken, chosen, identified, interested, […] stolen.
What is conspicuous here is the paucity of examples and the purely instantial nature of the description. No attempt is made to explain what it might be about the participles in either category that makes them behave as they do. Anyone with a question relating to the use of any of the many thousands of participles not included in Hewing's ultra-short list is bound to draw a blank, as would any student of mine desirous to understand why *the defined problem, *the bought equipment and *the arisen difficulties are unacceptable. Matters are made worse by the fact that the inclusion of 'used' in the first category is simply wrong. Ask any used car salesman. Hewing corrects this blunder in the second edition (Hewing, 2005: 138), replacing 'used' with 'included', something that would further confuse my engineering students, who need to be able to talk about such things as 'included sides', 'included angles' and 'included files' (standard terms in trigonometry and IT). The second edition amplifies the second category by adding 'alleged' and 'allocated', but again this would not help my students understand their mistake.
They might also be puzzled by Hewing's unreflected inclusion of 'interested' in both editions, because it blithely ignores the fact that a switch from its prenominal to its postnominal use is normally accompanied by a change of meaning (interested party vs. person interested).
Finding monolingual pedagogical grammars unhelpful and bilingual grammars (e.g. Beck & Gergel, 2014;König & Gast, 2007) insufficiently advanced to deal with this topic, I did the next best thing and ransacked the shelf of academic grammars at my local university library. The only one to address the question in hand was Quirk et al.'s Comprehensive Grammar of the English Language (1985), but since this weighty and scholarly tome extending to close on two thousand pages is sometimes regarded as the 'Bible' of English grammar, I was heartened.
Quirk et al. begin by distinguishing between active and passive participles and then observe (Quirk et al. 1985(Quirk et al. : 1327: […] the active [past participle] is rarely used in premodification. Contrast: the immigrant who has arrived BUT NOT: * the arrived immigrant The following are exceptional: a vanished treasure a retired teacher reduced/fallen/increased prices; risen costs (in the technical language of economics).
There are two problems with this. The first concerns the bold claim that use of the active participle in premodification is rare or exceptional. It took me no more than a minute or so to draw up the following list: a grown man a slipped disc a faded curtain a hung parliament an escaped convict an evaporated liquid a relaxed approach a collapsed lung a fulfilled life a troubled expression changed circumstances fragmented vision This does not mean that Quirk et al. are wrong in thinking that premodifying active participles are less common than premodifying passive particles -in fact I am sure they are right in this -but it does suggest that their claim that these are rare or exceptional is grossly exaggerated and therefore correspondingly unhelpful. The second problem concerns the seemingly erudite qualification 'in the technical language of economics' attached to the last example. The alleged exception invoked for technical language use is spurious because it matters not one jot whether reduced/fallen/increased or risen are used in an economic context or not. There is nothing unusual about the ordinary-language use of the following: a fallen tree the risen dough increased popularity This is disconcerting because it creates the impression that the authors are blustering their way through, making assertions without considering their consequences while simultaneously creating a false impression of analytic rigour. However, their account picks up when they observe: Premodification is somewhat more common when an active participle is modified by an adverb: the newly-arrived immigrant our recently-departed friend a well-read woman a soft-spoken person ? a recently-arisen problem This is incontrovertibly true, so we can have no quibble with the validity of the observation. What is disquieting is the authors' apparent lack of interest in what they are asserting. Surely anyone with a modicum of linguistic curiosity would find the phenomenon they have described intriguing. After all, what this means is that if you take a misplaced participial premodifier (as in, say, *a received letter) and expand on it as indicated (e.g. a recently received letter), it is suddenly no longer misplaced. This unexplained transformation is remarkable and, to a degree, counter-intuitive, because you might think that if something is misplaced, adding to it would only make matters worse. By failing to be surprised by this surprising phenomenon, the authors are to my mind missing a golden opportunity, as I shall endeavour to show later on. There seems to an assumption among grammarians that it is their job to describe and not to explain, but this creates a false dichotomy. Sometimes explanation is a necessary step on the road to description. However, for the moment, let us just acknowledge the correctness of the observation.
Quirk at al. now move on to consider the role of participles formed from passive verbs (Quirk et al., 1985(Quirk et al., : 1328: Most -ed participles have passive meaning, and only a few will easily admit the permanent reference that will permit premodifying use. We may contrast […]: The wanted man was last seen in Cambridge [the man goes on being wanted by the police] *The found purse was returned to its owner [the purse was found at a particular moment] But a lost purse is acceptable, because, although a purse is no longer regarded as 'found' after it has been retrieved, a purse will be regarded as 'lost' throughout the period of its disappearance. Again, there is an irritating lack of reflection from the outset. We have just spoken about the fact that amplifying a prenominal participle can make it acceptable where formerly it was not. Now we are suddenly confronted with the idea of permanent reference as a criterion for premodification. What is more, this criterion seems to have come out of nowhere, suddenly appearing on the scene without introduction as if its advent were long expected and its identity unquestionable. This tends to obscure the cognitive leap that precedes it. Amplifying a prenominal participle by adding an adverb can make it acceptable but it certainly does nothing to lend the resulting expression the permanent reference the authors are now invoking (there is nothing permanent about a recently received letter and certainly no increase in permanence between *a received letter and its acceptable amplification). What this means is that we seem to be dealing with two entirely different explanations for permissibility that are to some extent at variance with each other. Now, there is no a priori reason why this should not be the case, but it is a striking assumption that should give us pause, especially when it is introduced silently via the back door.
However, leaving that aside for the moment, let us take a closer look at this surreptitious newcomer. The first example was: The wanted man was last seen in Cambridge [the man goes on being wanted by the police] The commentary in square brackets may be materially correct but it is doubtful whether it constitutes a valid explanation. Consider: The wanted man was arrested last week and duly charged.
[the man goes on being wanted by the police??] Perhaps only the writers of Police Academy could imagine a scenario in which a suspect who has been apprehended and charged continues to be wanted by the police. So, what do we say here? Do we say: the man went on being wanted until he stopped being wanted (because he was arrested). This is doubtless true but is even less likely to constitute a credible explanation. What is more, it does not account for the unacceptability of: The *wanted outcome was a first-quarter profit of £10,000.
because it is surely equally true that the outcome in question was wanted until it stopped being wanted (because it was achieved or the first quarter over). The rest of the explanation fares no better: *The found purse was returned to its owner [the purse was found at a particular moment] But a lost purse is acceptable, because, although a purse is no longer regarded as 'found' after it has been retrieved, a purse will be regarded as 'lost' throughout the period of its disappearance.
Reading this is rather like watching your favourite cricket team trying to chase down a high score on a sticky wicket: you can't help admiring their determination but you just know they're not going to make it. The explanation is so laboured you feel even the authors cannot quite bring themselves to believe it. We can of course agree that the purse was found at a particular moment, but then it was presumably lost at a particular moment too. The lost purse's claim to permanence therefore has to rest on the fragile claim that it was regarded as lost while it was lost (which is questionable), but is not regarded as found once it has been found (which too is questionable). When lost property is returned to its rightful owner it is frequently because someone has found it and handed it in, whereupon it may well be classified as 'found' and remain so until claimed. But, as we all know only too well, many objects handed in within the day lie around the lost property office for months before they are finally auctioned off because their rightful owners failed to claim them, perhaps because they never realized that they were lost in the first place. This simply does not accord with the claim that 'lost property' contains some form of permanent reference absent in '*found property'.
But before we get too caught up in this particular example, it may be useful to test this all-important criterion on a few others. Consider: a hired car *a bought car a demolished house *a built house a delayed flight *a landed flight an unmade bed *a made bed According to the authors' argument, we would presumably have to say that a hired car is regarded as 'hired' for as long as it is hired (which we may concede), but a car that has been bought is not regarded as 'bought' once it has been bought (which is debatable). When the leasing agreement on my car ran out, I opted to buy it and now stubbornly insist on regarding it as 'bought'. The other pairings fare no better. I imagine the authors would want to argue that a house is regarded as demolished for as long as it is demolished. But the fact that we can say 'the demolished house was rebuilt' suggests that the concept of permanent reference here is trickier than it seems. The authors would presumably also want to argue that a house that has been built is no longer regarded as 'built' once it has been built, but this too is questionable. The supposedly permanent reference in 'delayed flight' is no less dubious. Flight schedules deliberately overestimate flying-time to allow for minor delays, so flights delayed on departure are not necessarily delayed on arrival. This aside, are we to believe that a flight that has landed is not regarded as 'landed' after it has landed? Every airport arrivals board in the world would seem to testify to the contrary. It is even harder to understand what greater degree of permanence sets unmade beds apart from beds that have been made. Which condition of the bed is more enduring surely depends on the household, but I would imagine that in most households the beds are made and remain made for considerably longer than they are left unmade.
Despite the obvious difficulties in applying it, the criterion of permanent or stable reference is surprisingly popular. Here is an entertaining example from another academic grammar (Radden & Dirven, 2007: 157) (one that was not on the shelf I ransacked): deposited money *withdrawn money together with the authors' (syntactically skewed) commentary: […] money which is deposited has stability, hence this transaction is expressed as [sic] premodification, while money that is withdrawn is no longer there, hence this transaction is expressed as [sic] postmodification. The authors seem unable or unwilling to contemplate the idea that money in the bank might not only be withdrawn but also embezzled or stolen, so it may be necessary to remind them that 'stolen money' or 'embezzled money' is also 'no longer there' (as they put it), but this does not seem to prevent premodification.
By now it should be clear that the much-vaunted criterion of permanent reference can scarcely bear the explanatory burden that is placed on it. A similar realization seems to have dawned on Quirk et al., who now hasten to add: Exceptions to the general rule suggests that the semantic and aspectual factors are more complicated than we have indicated […].
Indeed. A less generous view might be that the criterion is largely unworkable. But rather than admit this, the authors take a fresh run-up: The premodifying participle usually characterizes a type rather than an instance: a muttered reply is a type of reply, and a drawn sword a typical posture.
Here we now have a new criterion to supplement the previous one. The examples seem clear enough, though we may wonder what century someone is living in who regards a drawn sword as a typical posture. But let us test this criterion against our examples. Is a hired car a type of car? Possibly so, but probably not. By type of car we normally mean the make of car and/or the category of vehicle to which it belongs (saloon, hatchback, people carrier, van, SUV, pickup etc.). If you have a breakdown with a hired car (as I did recently) and the garage asks you what type of car is involved, you are hardly likely to say 'a hired car' (and I didn't). Such an answer would surely be regarded as somewhat eccentric and therefore potentially irritating to the interlocutor. It is almost not worth asking whether a demolished house is a type of house (I imagine not even the most desperate squatters would to assent to this), or a delayed flight a type of flight (though one or two airlines come to mind where one might be excused for thinking so), or an unmade bed a type of bed, or deposited money a type of money. The addition of typical does not help matters, because it raises the stakes rather than lowering them. To claim that a hired car, a demolished house, a delayed flight, an unmade bed and withdrawn money are typical examples of their respective kinds is surely even more implausible.
At this point Quirk at al. reiterate with regard to passive particles the point they made earlier in respect of active particles: [Otherwise unacceptable] participial phrases become acceptable when modified by adverbs […]: a recently sold car the above-mentioned article a well-built house a carefully described man Again, we are bound to concur and, again, we have cannot help noticing the authors' apparent lack of interest in the astonishing phenomenon they are describing. They conspicuously fail to ask the obvious question: 'why?'. Why does adding an adverb suddenly make an otherwise unacceptable construction acceptable?
So where does this leave us? We have been able to show that the accounts of the constraints operating in English participial premodification we have examined are largely unhelpful and often ill-considered. This may have helped to clear the ground, but it is leaves us wondering where to go from here.
When confusion reigns, it is sometimes a good idea to retreat to a point of reasonable certainty, just as Descartes once did. The Cartesian method has come in for considerable criticism but largely because the putative certainty to which Descartes retreated is often held to be illusory. The logic behind the method is surely unexceptionable and still useful in cases where a point of reasonable certainty can be established.
In the account of participial premodification provided by Quirk et al., there was one point we felt bound to agree with, namely the observation that an otherwise unacceptable participial premodifier (as in '*the received letter') is suddenly rendered acceptable by adding an adverb ('the recently received letter'). I have already suggested that the authors' refusal to interrogate this remarkable phenomenon is unwise, because interrogating it may yield a vital clue.
What seems to disturb us about '*a received letter' is that the participle tells us nothing of any consequence and is therefore apt to puzzle us. Letters are sent and received -such is the nature of letters -so merely describing a letter as 'received' is strangely unhelpful and raises more questions than it answers (when? from whom? under what circumstances? etc.). The addition of an adverb increases the semantic weight of the premodifying phrase and provides the otherwise inconsequential participle with a rationale: 'a recently received letter' tells us something of consequence by setting the letter in question apart from letters received earlier. What this suggests is that the prenominal position of the participle is a privileged one that can only be occupied by a term significant enough to deserve it. You might compare it to a VIP parking space that can only be occupied by those deemed to merit it.
Interestingly, the idea that a premodifying participial phrase has to have a certain semantic weight or semantic salience (as I think I would prefer to call it) for the participle to deserve its privileged prenominal position offers an intuitively more plausible explanation for the 'lost property'/'*found property' conundrum than the criterion of permanent reference. 'Lost property' commands our attention (something is lost and needs to be found) in a way that '*found property' does not. The same is true of the other examples we considered: 'a hired car' vs. 'a *bought car' (a hired car is only temporarily in my care and needs to be returned at the appointed time); 'a demolished house' vs. 'a built house (a demolished house is uninhabitable, in fact not even a house at all any more); 'delayed flight' vs. '*landed flight' (all flights are supposed to start and land but not all flights are supposed to be delayed); 'an unmade bed' vs. 'a *made bed' (an unmade bed needs to be made); 'deposited money' vs. '*withdrawn money' (deposited money has a distinct financial status whereas almost all money has been withdrawn from somewhere at some time). We can go further. Consider: a broken promise *a kept promise incurred losses *gained profits Promises are meant to be kept so promises that are kept are to that extent unremarkable while broken promises are a cause for concern. Not all losses are incurred -some are the result of forces beyond our control -but all profits are gained.
We may even be able to use the criterion of semantic salience to account for the following: the required result *the needed result the desired outcome *the wanted outcome These cases are initially somewhat baffling, since in each pairing the acceptable premodifier and the unacceptable one would normally be considered virtually synonymous. What difference there is would seem to be mainly a matter of register: required/desired are high-register while needed/wanted are lowregister. But how can register influence the acceptability of a participial premodifier? Well, what we could say here is that the formality of required/desired implicitly relates the phrases in question to something we presume to have been formally defined elsewhere (in a business plan or strategy document or contract, etc.). When used in a suitable context, the formal term therefore acquires a semantic salience absent in the informal term, which tends to raise more questions than it answers (needed by whom? wanted by whom?). Where the apparently informal term is acceptable ('wanted person'), it seems to be because it is the formally accepted term its own particular context. It remains to be seen whether the criterion of semantic salience that I have proposed is robust enough to be both universally applicable and didactically useful. I have my reservations because I am not sure how well it can deal with a case like this: a burgled house *a robbed bank Does it make sense to say that the former has greater semantic salience than the latter? Of course, you could argue that the image conjured up by 'a burgled house' (broken locks, smashed windows, ransacked cupboards, overturned drawers) is more dramatic than image we have of a bank that has been robbed, because although a bank robbery may be a fairly dramatic event it probably leaves the bank looking pretty much as it did before. But this may be stretching things a little. We risk batting on the same sticky wicket as Quirk at al. trying to convince us that lost property has greater permanence than property that has been found. Another challenge will of course be to provide some sort of account of the degree of semantic salience needed to justify premodification. This presumably cannot be achieved long-winded theoretical explanation but needs to be illustrated by means of examples; however, not arbitrary examples like Hewing's but examples selected to shed light on the critical threshold.

Conclusion
Nonetheless, I think some meaningful progress has been made. Instead of disparate criteria that are to some extent at variance with each other, our retreat to Cartesian certainty has yielded a single criterion that not only provides a coherent explanatory model (thereby satisfying not only Descartes but also William of Ockham, whose famous razor urges us to prefer a single explanation) but also explains many of the cases we have considered better than Quirk at al. have been able to. What is more, I now have something to throw at my students whose continual misuse of participial premodifiers has (I am ashamed to admit) sometimes left me itching for a suitable projectile. If we have learnt anything here, it is that premodification is not open to participles whose meaning is trivial or banal or inconsequential. We do not say '*a built house' because almost all houses that we are likely to talk about have been built (it is usually only architects who talk about unbuilt houses). We do not say '*a bought car' because we regard ownership as the standard from which alternatives (hired cars) need to be distinguished. We do not say 'a kept promise' because promises are meant to be kept -that is whole point of promises -so only a promise that is broken commands our attention. This notion can be applied to: *the defined problem *the bought equipment *the arisen difficulties *the overcome obstacles *the conducted experiment *the obtained results *the found solutions What I can now say to my students is that the meaning expressed by the participle in these cases is too inconsequential to warrant privileged prenominal use: it lies in the nature of solutions that they are found; it lies in the nature of results that they are obtained; it lies in the nature of experiments that they are conducted, and so on. The triviality is compounded by the fact that what the students are invariably doing here is using the participial phrases for purely anaphoric purposes. By '*the defined problem' they mean 'the problem (whose definition) we have just been talking about'; by '*the bought equipment' they mean 'the equipment (whose purchase) we have alluded to', and so on.
Having played my hand, I would like to return for a moment to the provocative question implicit in my title: 'When grammarians fail us' … what then? If I am to risk a straightforward answer, I think it would have to be: 'Have a go yourself. Who knows, you may do better'. In the present case, I am bound to add, it would be hard to do worse. But is it enough simply to quibble with grammarians? Are there perhaps not questions of a grammatical nature that cannot be fully dealt with within the confines of a grammar book? This in turn brings me back to Dixon, whom I mentioned at the outset. Dixon criticizes lexicographers for paying insufficient attention to grammar. If he is right in this, it seems reasonable to ask whether dictionaries might not also have an important role to play in dealing with the issue we have been examining. After all, dictionaries could tell us which participles allow straightforward premodification and which do not, giving examples of usage and making the necessary semantic distinctions (e.g. 'wanted person' but not '*wanted outcome'). The idea behind this is not that grammarians should offload the problem on to lexicographers but that both might cooperate in drawing a fuller picture and in providing the guidance we need and expect.
Grammar cannot be neatly separated from lexicology just as it cannot be neatly separated from syntax, semantics, morphology, phonetics and pragmatics. It is important to reflect this by considering ways in which the various disciplines can profitably interact.