Is wrong article use a matter of pronunciation or grammar?

Question

I was in a discussion with someone, where they described my wrong use of an article as a "mispronunciation". I argued it was rather a matter of grammar, as I did pronounce the article correctly, but it was the wrong one. I should say now that this isn't the English language we're talking about, we were talking Norwegian at the time.

First of all, I did not use the wrong tense, i.e. using indefinite articles instead of definite, which would definitely be a matter of grammar. My offense is more in the lines of using "a" instead of "an", but not entirely. As you probably know, whether one uses "a" or "an" is based on whether the word in question phonetically starts with a vowel or a consonant. If it starts with a vowel, an "n" is put after the base "a" to make it more appeasing to utter and hear. This is a concept that I believe is quite wide-spread (at least in Indo-European languages), with an example of this in Spanish; it being el agua instead of la agua.

And so, with these languages, I could see that one would call it a fault of pronunciation to use "a" instead of "an", as the very distinction exists for the purposes of refining the "sound of English". But in my case, I was speaking Norwegian, of where the distinction between the articles is not based in any such logic. To be honest, I do not believe there is any logic to it, and if it was, it has disappeared in the evolution of the language into weak and blurry silhouettes of patterns. It doesn't inherently, or according to some establishment of logic, sound better to use the correct articles in Norwegian. Rather, it sounds better because we have been "taught" to use those articles, and we've heard them been used that way many times, and so a deviation from that sounds bad.

I'm not claiming that a foreigner will necessarily find "an" to sound better than "a" in front of a word starting on a vowel, but there is a logic behind it based in phonetics. In Norwegian, there are simply rules saying that "a word constructed such and such shall use those articles respectively". These rules are littered with exceptions, but they exist there. But they're not based in any logic, at least not one present today. And so, it is simply arbitrary grammar in my head, and to use the incorrect article would be a matter of exactly that, grammar. But perhaps it isn't that, but instead a matter of pronunciation. Or maybe both. Or neither. Perhaps there isn't an objective answer at all?

Answer 1

The question draws a false albeit common dichotomy between grammar and pronunciation. Grammar includes the facts that in English we say "I saw your uncle" but not *"I your uncle saw", which is about order of words. Or we cannot say *"Took the car", because it lacks required words. Or, we cannot say *"Me likes she" or *"I have see you" because one is using the wrong form of words in particular positions / syntactic functions.

In the case of "a" versus "an", there is a particular morpheme (the indefinite article) whose distribution is sensitive to the phonological properties of the following word. This feature where morphemes have multiple realizations is known as "allomorphy". Unfortunately, naming the phenomenon does not explain what it is. There are two theories of a/an. Theory 1: /n/ deletes by a phonological rule in the indefinite article when the next word begins with a consonant (one can also state this as n-insertion). Theory 2: select the morpheme /an/ before a vowel when inserting the correct variant of the indefinite article, as a rule of morphology. Both accounts rely on the notion of "rule of grammar", they just invoke different kinds aspects of grammar (morphology vs. phonology). The various forms of the past tense in Norwegian with -de, -et etc. are similarly examples, where phonological properties are also relevant in determining which form is used.

I suppose if one said [fi] for "we" or [hɛn sɔ:t] for "He sat", those would be errors of pronunciation. There are no rules in any part of the grammar where you can say "You violated this rule". Something like [fi finner finnbiff feldi fint] would be a kind of articulatory error, thus "pronunciation". Saying ei glass; en glass is not plausibly a pronunciation error; it would be most likely a lexical error, not grammatical or pronunciation (okay, it's highly implausible since everybody know the gender of "glass"). Something like [en gɔtt røtt egg er fint] probably has to be treated as an error of grammar (it's not the person has the wrong gender for "egg"), and [e gɔtt røtt egg er fint] would most likely be a pronunciation error.

Answer 2

You may be making some confusion, and likely so is the other person. In English, a vs an (and the pronounced in its two different ways, too, which is similar but just not reflected in spelling) is a completely different distinction from Norwegian en vs et (vs ei) or the corresponding suffixes for the definite article: in English, as you say, it has to do with pronunciation (an is older than a, and has the same origin as one; the eventually disappeared before consonants); in Norwegian, the article is decided based on grammatical gender.

Depending on your dialect, there is a masculine vs feminine vs neuter, or just a common vs neuter distinction. It's called gender in both cases, and goes all the way back to Proto-Indo-European. This distinction disappeared in modern English, but Old English had it, and it is preserved, for example, in German (think der, die, das) and in Romance languages, which have mostly lost the neuter and only kept masculine and feminine (think el vs la; the "agua* case is special, as it's a feminine noun and those usually take la, but indeed, in that case it takes el because of pronunciation matters, i.e. it starts with a stressed vowel).

In all Nordic languages, the distinction is kept, although the standard form of Swedish and Danish merge masculine and feminine into a "common" gender, which is optional in Norwegian bokmål. Some dialects in all three languages may preserve masculine and feminine separately. Icelandic also preserves the full distinction.

So to conclude: it's definitely a grammatical distinction in Norwegian. A noun always keeps its gender, which determines which articles to use. In English, you can call it a matter of pronunciation, in part because unlike the gender system, the article you will use depends on a characteristic of the following word, which is not necessarily the noun; but, really, it's not the kind of thing that phonology tends to get into, not at this late a stage (when speakers saying a no longer think anything like that they're "just sometimes saying an in a lazy way in front of some words", but instead have grammaticalized the distinction), so I'd call that grammatical too.