One of the more difficult things to do in grammar instruction is to define what a sentence is. It is almost as difficult as defining what a word is. No kidding.
Words seem self evident. They are units of language that are made up of letters and have an identifiable meaning or definition that can usually be found in a dictionary. In speech they can be pronounced as a single unit, while in print they have spaces on both sides. That seems simple enough.
We could keep this simple and exemplify this with a solid word, such as the word “establish,” and proceed to find it in the dictionary. Then we would find that even dictionary.com has eight definitions for the word. All of them are the same part of speech, so this is not really that much of a problem. (I can imagine that the OED has a few extra definitions, but that is a silly book we should not trouble ourselves with at this point.) But if I look at how we defined the parameters of a word as having an identifiable meaning, would “establish” be one word or eight? I will try not to think about that right now.
Instead I will move on to another issue brought up by the modification of words with suffices. (Is that supposes to be “suffices” or “suffixes”? And would that count as one word or two? They seem to be two different words but they have the same exact definition. No matter. I should be writing and not thinking.) I could, for example, add an “es” to the end of “establish” and make it “establishes.” Does this count as a different word or a modification? Probably a modification, eh? How about if I slap a suffix on the end and change it from the verb “establish” to the noun “establishment.” Seems good. How about if I add an “s” onto the end and make it “establishments”? Not really? How about if I slap a prefix onto the “establishment” and create “antiestablishment.” Or “antiestablishmentarian.” And then “antiestablishmentarianism.” And we should not forget “antiestablishmentarianisms.”
Our most basic unit for grammatical instruction seems hopelessly undefinable in any practical sense. I am sure a good linguist could go through all this silliness and make sense of it with precise definitions of concepts only a linguist would be familiar with. The problem is that a linguist’s precision would quickly become silliness for most students.
We run into a similar problem with the definition of a sentence. The most common definition I have heard is probably the one that has created the most confusion. Most students come into class with the definition they picked up in junior high or high school: “a sentence is a complete thought.”
This definition tells students very little since a “complete thought” can be as simple as a one word answer to a question. If someone asks you “When are we going to Disneyland?”, an answer of “Tuesday” would be perfectly complete since nothing else is needed to answer the question. A complete thought could also answer the question “What is a sentence?” but that question seems to be demanding more than a one word answer at this point since I am nowhere near completing this thought.
The “complete thought” definition becomes mildly useful, however, when students try to determine if a dependent clause can be punctuated as a sentence. The dependent clause “When Claude was waiting in line at the Matterhorn” sounds as if something is missing. Most student say they want to know what happened when Claude was waiting in line.
It seems many people have this vague sense that something is missing when there is no independent clause. Asking about what happened when Claude was waiting in line forces the students to come up with an independent clause that would change the fragment into a sentence. It would seem that the “complete thought” people might have a point about fragments. But, thankfully, they do not. A simple erasure of the subordinating conjunction “when” would change things entirely. The stripped down clause would read “Claude was waiting in line at the Matterhorn,” which most people would probably consider to be a complete thought, but this transformation happened not because something was completed, but because something was eliminated.
I do understand that linguists can break this stuff down far more effectively than I can and are able to provide more precise definitions of words and sentences. I get that. These linguistic definitions are not useful for someone who wants to write better for a job and does not care much for discussing the shortcomings of Chomsky’s universal grammar in light of the fMRI studies published in Volume 114 of Brain and Language. The problem is that when we simplify things for grammar instruction, we often create problems for our students. That is usually not the goal of simplification.
But instead of doing away with our rules and definitions because of their short comings, we need to make it clear that many of the terms we work with are good only within a certain context and are not so much rules as they are conventions. Students who learn grammar in college just want to write better and have no use for the particulars of language. In a sense, our students are learning to become mechanics and need to know and need to know where to pour the oil; they are not learning to become mechanical engineers who need to calculate velocity using the Bernoulli theorem.