Researching language online: whom do you trust?

research

This short guide is addressed to teachers, trainee teachers and teacher trainers, not learners of English.

Frequently, teachers are encouraged (or even compelled) to do some research into the language they are teaching or writing about by accessing various website resources.
Because native speakers of any language can, with some justification, claim to be experts in the language, and also because the language they speak is usually of passing interest to them, there are hundreds, possible thousands, of websites designed, written and maintained by well-meaning people with the intention of helping people to understand the structural and functional characteristics of English.
Moreover, because English is the most widely used language on earth, there are naturally many thousands of people who are interested in its nature and they, too, have a right to claim some expertise, even if the language is not their first.
That's all very well, of course.

The problem comes when what people say on their friendly and sometimes well-designed websites is wrong.
In what follows, we have withheld the sources of some citations to spare the blushes of their authors. They are pretty easy to track down, but why would you trouble to do so?

So, what's the problem?

The problem is, whom do you trust?

Let's say, for example, that you are looking for some reliable information about adverbs and prepositions.
These are major word classes, not just in English, one of which is an open class to which we can make additions (adverbs) and the other of which is a closed-system class of words (prepositions) to which it is very rare for additions to be made.
An additional distinction is that adverbs qualify verbs in some way (or other adverbs and adjectives) and prepositions form the heads of prepositional phrases and are followed by a complement or, in more functional analyses, an object.
It should, you would probably agree, be a simple matter to distinguish between them.

Unfortunately, there are people writing or contributing to websites who have a shaky grasp of the difference between adverbs and prepositions.
Here's a selection or citations from websites. Can you identify the error and correct it?
Click on the to reveal an answer.

Phrasal verbs change meaning based on the preposition that follows them. One example which follows is: Could anyone with information about this crime please come forward.	Well, actually, phrasal verbs are not formed with prepositions at all. Those are called prepositional verbs. Phrasal verbs are formed with adverb particles. In the example, of course, the word forward is an adverb, not a preposition. In fact, it is a slightly rare case of a word of this kind which can never be a preposition (although it can be an adjective). The expression come forward cannot be described as a verb plus a preposition or as a phrasal verb (the sentence also needs a question mark, just incidentally). What it is is a verb followed in the usual way by a modifying adverb. It's not mysterious.
Preposition OUT is opposite of IN and used to show movement away from the inside of a place or container. Examples which follow include: Your brother was out when I came by to see him. Government forces have driven the rebels out of the eastern district. Are you going out tonight? We're heading out at seven, so don't be late.	It hard to unravel this nonsense which is designed, apparently, to help people with phrasal verb use. Firstly, of course, out is usually an adverb and out of is the preposition. Out of can only function as a preposition but out can be either an adverb or a preposition (or an adjective or a verb, incidentally). In the first example, the word out is an adverb, some might argue that it's an adjective, but what it is most certainly not is a preposition. In the second example, the preposition is out of, not out and that is followed by its object or complement the eastern district. It's an example of a verb followed by a modifying prepositional phrase. In the third and fourth examples, we have out as an adverb again and in the last case the adverb is followed by a prepositional phrase. In none of the examples (and there are many of them which follow) is the word out used as a preposition on its own although it is possible to use out in limited circumstances as a preposition.
The following 'rule' is given on a website ostensibly intended to help learners with prepositions for the TOEFL examination: A preposition is a word which, with the following noun or pronoun, forms a phrase, and shows the relation of its object to the word whose meaning the phrase modifies. One example which follows is: The boys studied until they were tired out.	The rule is, as you see, incomprehensible and even if it were not, it's probably wrong. In the example, the word out is an adverb.

There is no argument that it is sometimes hard to tell an adverb from a preposition, especially if the prepositional complement or object is ellipted, but that does not mean we shouldn't try.

The problem is compounded, additionally, on websites which cannot distinguish between a verb modified by an adverb in the normal way and ones in which the adverb combines with the verb to form a new meaning. Nor can many distinguish a prepositional phrase following a verb from an adverb forming a phrasal verb. We get, therefore, all of the following described as containing phrasal verbs:

    The crane picked up the entire house.
    They tried to come in through the back door, but it was locked.
    It was so hot that I had to take off my shirt.
    Stand up when speaking in class, please.
    Someone broke into my car last night and stole the stereo.
    Sally was about to get on the plane, but she turned around when someone called her name.
    I need to get rid of her.
    We let our lovely dog in the house every morning.
    We got on the bus at the usual stop.

None of those sentences contains examples of phrasal verbs. Most have verbs modified by adverbs in the normal way and others are, in fact, verbs with modifying prepositional phrases.

Here are some more examples of errors of analysis with which the web is infested. You can correct them and then click on the to see if we agree.

Relative clauses are clauses starting with the relative pronouns who, that, which, whose, where, when.	No. Of that list, only the first four are relative pronouns (and one or two are missing). The other two are relative adverbs and they work differently. This is a common mistake because the grammar is superficially similar.
Modal verbs help us understand more about the verb in question. They give us hints on the possibility of something happening (can, should, etc.) or time (has, did, was, etc.)	Half right but poorly explained. In fact, verbs such as has, did and was are not modal auxiliary verbs at all. They are primary auxiliary verbs and contain no hint of modality. The clumsy expression They give us hints ... is about 20% right.
Adverbs of Time at the same time as for a long time frequently from time to time in a few minutes in the mornings last week	Some of these are actually adverbs (such as frequently) but most are prepositional or noun phrases. They may be adverbials in certain circumstances, but they are not adverbs.
Collective nouns a bottle of milk a herd of cattle a hack of smokers a cup of tea a staff of employees	There are two obvious problems here. a bottle of milk is not a collective noun, it is the opposite, a partitive expression. Collective nouns refer to the mass made up of individuals, partitive expressions refer to individuals from the mass. The second problem is the failure to distinguish between a collective noun such as staff or choir which does not need the of expression and an assemblage noun which usually does. We have, therefore, no need to insert of employees after staff or of singers after choir but we do need to insert the word cattle after an assemblage nouns like herd. The final problem is that there is no such thing as a hack of smokers. It's made-up nonsense.
We use some and any with uncountable nouns and plural nouns. The general rule is that you use “some” in positive sentences and “any” in negative sentences and questions. “I have some ideas.” “I don’t have any ideas.” “Do you have any ideas?” However, we can also use “some” in questions. “Would you like some tea?” (I expect the answer to be “Yes”.) When we use some in a question, we limit what we are offering the other person.	There are obvious problems with this: This will mean that all the following are wrong: Anybody can come in Anything you can do would help I don't know some of these people Do you know something about this? What if someone calls? She denied stealing any money And they aren't. It's a quasi-rule and no real help at all. The second problem is that "Would you like some tea" is not a question, it's an offer. What the speaker expects the answer to be is a mystery. The third problem is the last bit about limits. This seems to have been made up because it is wrong.
Words like deer and furniture do not have plurals.	This is surprisingly elementary. The word deer, in common with some others for animals in particular (but not solely) has what is known as a zero plural. That is to say, it is unmarked in the plural. It has a plural, of course, because we can say, for example: Three deer are in the garden Furniture, on the other hand, is a mass noun and mass nouns have no plural. There is a very simple difference between no plural and a zero marked plural.
Real vs. Really Real (adjective) describes things, really (adverb) describes actions. He is a real hero. He is really heroic.	This is another surprisingly elementary error (it comes from someone keen to sell you his books on English grammar and usage, incidentally). Yes, the word real is an adjective and the word really is an adverb. So far, well done. The problems come next: Adjectives describe things (true) but adverbs do not describe verbs, they modify them. They also modify adjectives and other adverbs and sometimes prepositional phrases. What we actually have here is an adverb modifying the adjective heroic. It's called an intensifier. It's hard to see how the word really describes anything and impossible to see how the word heroic qualifies as an action.

The examples above are all in really very simple areas of English grammar and structure. The list could be extended almost indefinitely if we include trickier areas of grammar and phonology.
Fortunately, however, most of the authors of the websites in question have not had the courage to venture into more esoteric areas.

The worst by quite a long way are two sites written to 'help' people prepare for and take a CELTA course.

Here's an example of what is meant taken from http://passthecelta.com/course-skills/parts-of-speech/. All headings here are taken verbatim from the site and are in this type face.
You can correct them and then click on the to see if we agree.

The site lists what it calls the basic parts of speech and that list is: subject, object, preposition, adverb, verb, adjective, noun (countable), noun (uncountable but that term is left out), auxiliary verb, modal verb, pronoun, article, conjunction.	Here are the problems with the beginning: That is not a list of the word classes in English or any other language (or even the old-fashioned parts of speech). The only word classes proper in that list are: preposition, adverb, verb, adjective, noun, pronoun and conjunction and the list ignores determiners and interjections. The other categories refer either to a sub-class, such as auxiliary verb (which should be called primary auxiliary verb), modal verb (which should be called modal auxiliary verb) and articles (which form a subset of determiners). Subjects and objects are not, of course, parts of speech or word classes at all. They are, technically, arguments and their function is to do with syntax not vocabulary. The site lists, incidentally, only coordinators under the examples of conjunctions and despairingly sends you off to another confused, inaccurate and confusing site should you feel you are not confused enough.
Unlike native speakers, language learners need to construct their sentences carefully, paying attention to which words perform which functions (parts of speech) as they interact in the phrase or sentence.	This is slightly confused and misleading. One of the guiding principles of teaching a language is to stop this over-monitoring behaviour by learners. Other problems include: It is certainly true that learners often do need to pay attention to word class (the correct term for the dear old fashioned 'parts of speech' incidentally) but so do native speakers, particularly the one who wrote this site. The assumption is implicit here, too, that one is either a native speaker or not and that disguises all the intermediate steps between being a non-speaker of a language and a completely fluent one. A word's grammatical function (if that is what is intended by the term here) is not the same thing as its word class. A noun, for example, can function grammatically as a direct or indirect object, an adverbial or a subject depending on its syntactical properties in the context but that's a slightly technical point. It is also imperfectly clear how one can *interact in the phrase or sentence*.
I will pass my CELTA course. subj + aux + v + pron + obj	This is an attempt to see how a sentence can be parsed into its various components. Unfortunately, it's wrong. Here's why: The first problem is that the first word is labelled the subject. It is, of course, but subjects do not form a word class, leave alone a part of speech. It is in fact a simple subject-case pronoun. And pronouns are one of the main word classes. Secondly, the word my is not a pronoun. It stands for nothing (which is what pronouns do). What it is is a possessive determiner. If you call it a pronoun, you are misleading the learners unforgivably. The correct word class to assign the word to is determiner. If you want to be very old fashioned, you could even call it a possessive adjective, but it still is not a pronoun. Lastly, we have another non-word class classification (obj). Objects are also not word classes, they are words, phrases or clauses which perform a certain grammatical function in a sentence. For example, I want what you gave Mary contains an object of the verb want but it is what you gave Mary and that is a clause not an example of a word class.
All full sentences must have a subject.	Not true, even if one can figure out what the writer means by a 'full sentence'. Try: Go into the kitchen and ask your mother for example. That is a sentence (called a compound sentence) but it contains no subjects.
Verbs also have a form called the Past Participle. These are when we use verbs as an adjective to describe something. The book was TORN. The necklace had been GIVEN as a birthday present.	There is a form called a past participle and both the examples in upper case in that assertion are correctly identified. Well done so far. Unfortunately, the rest of the assertion is wrong: Past participle forms can be used as adjectives (as in the first example with torn) but that is not their main function. When a word is used as an adjective, it is an adjective. Suggesting that a verb remains a verb even when it has been converted to another word class is deeply confused and confusing. In the second example, the participle given is not an adjective at all, it's a non-finite verb form making part of a passive voice sentence. (We are not here to discuss the awful concord issues in These are when we use verbs as an adjective.)
Adjective (adj.) Adjectives are really easy to remember. They simply describe or modify something. large, red, angry, beautiful, essential, tasty, Korean, leather etc. The LEATHER chair looked WORN but EXPENSIVE. My KOREAN teacher turns RED when he is ANGRY.	Half marks. There is a difference between describing and modifying and a difference, too, between describing and classifying. In this little mishmash of unlikely examples, we actually do have four adjectives: worn, expensive, red and and angry. All of them are used predicatively by being linked to the noun they describe with a verb (look, turn and be in this case). Not all adjectives can be used that way and some must be used that way. The first of those, worn, is, by the way, a participle adjective and belonged in the last example. The other two modifiers (leather and Korean) are not central adjectives at all, they are classifiers which tell you the category of the noun rather than describing it. Unless you know the difference between a classifier and a central adjective you won't know why The chair was worner He was less Korean *It was very leather are wrong.

The website owns up to:
This website represents the personal opinions and shared advice of a single individual.
It is not affiliated, approved or endorsed by University of Cambridge ELA.
which, considering its error-strewn content, is probably just as well.

You are, of course, at liberty to use that website or any others you may come across in preparation for an initial or any other training course and you could use it a resource when planning and preparing but you'd be unwise to do so.
You could even invest £59 in a set of materials which were all written by someone who passed the CELTA with the highest possible grade! (unnecessary punctuation in the original). That means, of course, someone who has completed 120 hours of basic training in a profession in which others have spent decades and are still learning.
Hopefully, the commercial material has been written with slightly more care but we are not about to spend £59 to find out.

Lest you think that this is the only example of a site written by the inexperienced and ignorant for the inexperienced and ignorant you should know that in researching this short guide six other similarly confused and unhelpful sites were investigated. It would try your patience too far to include all that was discovered and, anyway, the point has been made.

On this site, a number of the areas of language analysis carry website warnings and many concern the sorts of issues we have exemplified here.

So, how do we decide whether to trust a site?

The first thing to do is to get some idea of whom the site is intended to help. There are four sorts of site (although some straddle the boundaries and some are almost impossible to characterise with any confidence).

Sites designed to help learners of English.
Many of these are constructed by commercial organisations who want to use them as a free hook in the hope that visitors will go on to access the businesses' main sites and book a course or buy some materials.
Others are written by people who would like to earn some money from teaching online. They, too, are often obvious hooks intended to encourage people to sign up for courses (and pay for them).
They have the following main characteristics:
1. They are often written by teachers on the staff of the organisations who are under-trained or otherwise somewhat ignorant.
2. They try to simplify the area because they know that their object is pedagogic rather than to do with training teachers. There is nothing wrong with that: many text-based grammar books do the same thing but it is usually clear what the intended audience is.
  If the audience is learners of English, you will find that grammar guides in particular are likely to be simplified to the point of inaccuracy. It is very difficult to remain both accessible and accurate. That is not a criticism; it arises from the intentions of the authors.
3. They are often wrong.
4. They are usually hurriedly put together and contain quite numerous typographical errors.
Sites designed by publishers for teachers of English.
It has not escaped the notice of major ELT publishers that teachers are often quite influential when language teaching organisations come to enhancing or replacing their materials banks. Naturally, establishing a resource for teachers is a good way of raising the profile of certain materials and their publishers. Most major materials publishers now have websites which appear quite independent and unaffiliated but are, in fact, paid for and maintained by the commercial organisation which lies behind them. Some are more honest and open about this than others.
They are variably trustworthy, naturally.
Sites designed to be areas where speakers of English can discuss their ideas about how the language works.
These sites are usually places where people can post and answer questions about the language and are open to all. Such sites are not primarily intended for people teaching English or training teachers of the language. They have the following characteristics:
1. The topics are randomly organised because the structure of the site depends not on any consistent approach to analysis but on the questions people pose and the issues they raise.
2. The answers which are provided are supplied by self-selecting people who feel they have some knowledge to impart. Sometimes, this is justified, often it is just a re-hash of half remembered and poorly digested information.
Sites designed for training teachers or other people studying language seriously.
These sites are often written by people working in higher or further education as a supplement to a face-to-face or online course in linguistics, pedagogy or applied linguistics.
Others are constructed by large non-profit organisations who have the resources and expertise to be accurate and helpful.
A few have been written by real experts in the language and teaching who are using the site to promote their own published material or just because they want to be helpful (or both).
They are:
1. Usually quite trustworthy because they are written by academics with a reputation to uphold and an understanding of the pitfalls to avoid.
2. Often accessibly organised into topic areas and some have an internal consistency which helps people to build on their knowledge as they work or read through the materials.

Scepticism

If you are reading this, you are probably a teacher of English or a trainer so the first kind of site is not for you. Some of the materials may be helpful when preparing a particular lesson (because that's the audience they are aimed at) but they are worse than useless if your concern is to analyse the language and research the area.
The issue here is that you have to check what you are being told very carefully (usually by cross-checking with a proper grammar book or a trustworthy site).
Do not trust them.

Sites designed by publishers can be extremely helpful if you are trying to locate a particular set of materials or a worksheet to insert (suitably adapted) into a lesson. They are less helpful when it comes to language analysis because analysis doesn't sell books.
This second kind of site also requires you to cross-check the information you find.

The third kind may be quite interesting and sometimes can alert you to information about the language of which you were previously unaware. However, their organisational shortcomings of this sort of sites may mean that you can't easily find what you are looking for and when you do, that you find it isn't what you wanted.
Sometimes, as was noted here, the responses to people's questions and the issues raised are accurate, accessible and comprehensive as well as written by people who know what they are writing about. Unfortunately, that is often not the case so a little scepticism is in order.

The fourth kind of site is the sort you can probably trust.
This site likes to think it is one of these.