There's a system I use to make memorizing different rules and irregularities, particularly in word inflections, much easier.
Every word, every sentence, every utterance in general has, on the phonological level, two representations: the underlying, or "phonemic" (how it's stored in speakers' heads) and the explicit, or "phonetic" (how it's actually spoken out loud). See Phones vs. Phonemes for more on this.
This can also be applied in a deeper, more abstract sense to take seemingly irregular inflections and make them regular. For instance, many verbs in Spanish have "stem-changing vowels". The word pensar, "think", when conjugated in its forms for the present tense, goes:
(yo) pienso (tú) piensas (él, ella) piensa (nosotros) pensamos (ellos, ellas) piensan
Note the apparent irregularity: the /e/ in /pensar/ turns into /ie/ in some of the forms! However, this is actually a very simple rule: when the /e/ is in a stressed syllable, it becomes /ie/. Otherwise, it remains /e/. All the forms with /ie/ have that syllable stressed, but /pensamos/ is stressed on the second syllable, and so this rule doesn't apply to the /e/ in that form of the word.
Now, if you approach storing this verb in your mental lexicon in the right way, this is just one extra bit of information to tag the verb with. When memorizing the word, instead of only memorizing "pensar", memorize "pensar, e -> ie". Since this is a very common feature for Spanish verbs to have, it becomes its own regular pattern! Other words like "querer, e -> ie" (since it does the same thing), "mover, o -> ue", etc. all have the same rule. Some verbs have "o -> ue" instead of "e -> ie", but that doesn't matter; the rule doesn't discriminate between vowels. All it says is that the first vowel becomes the second one when that syllable is stressed.
So, we've taken a verbal pattern that often causes learners of Spanish headaches and turned it into another feature that the verb has. A full mental entry for the word "pensar" could look like:
pensar - suffix vowel /a/ - stressed /e/ -> /ie/
It's just another bit of information to tag the verb with. With this system, if we want the first person singular of the verb, we do it in the following steps:
- Take the stem form of pensar, /pens/
- Add the first person singular suffix, /o/: /penso/
- Apply the stress rule: /pienso/
And that's it. The only thing you have to memorize is the one extra "e -> ie" rule, instead of a bunch of irregular forms.
The French verb vouloir, "want", can be analyzed in a similar way. Its conjugations go:
(je) veux (tu) veux (il, elle) veut (nous) voulons (vous) voulez (ils, elles) veulent
French spelling only serves to make things more confusing, so here's the same table but in phonemic transcriptions:
1st. sing.: /vœ/ 2nd. sing.: /vœ/ 3rd. sing.: /vœ/ 1st. plur.: /vulɔ̃/ 2nd. plur.: /vule/ 3rd. plur.: /vœl/
What are some observations we can make about this? Keep in mind that in French, the last syllable of a word is always stressed.
- Stressed syllables all contain the vowel /œ/
- Unstressed syllables contain the vowel /u/
- The plural forms all have a /l/
Therefore, we can take "vouloir" and store it like this:
"vouloir" /vulwar/ - root: - stressed (default): /vœ/ - unstressed: /vu/ - plural indicator: /l/
So, next time we want to construct the 1st person plural form (nous) of this verb, we can do it in the following steps:
- Retrieve "vouloir" from the mental lexicon.
- Take /vœ/, the default form
- Add the plural suffix /l/
- Add the 1st person plural suffix /ɔ̃/
- This triggers the unstressed form of the root (since we've added a new syllable, which is now stressed), giving us /vulɔ̃/
- Generate the final phonetic form: [vulɔ̃]
/vœ/base + /l/plural + /ɔ̃/1st = /vœl/ + /ɔ̃/stressed -> /vulɔ̃/ ->filter [vulɔ̃]
This may seem needlessly complicated, but the best part is that this exact template works for many other verbs, so in memorizing it once, you're actually making a bunch of other verbs easier, too.
Note that the "filter" step is simply the process of turning phonemes into phones, as discussed in Phones vs. Phonemes. In this case, there aren't any complicated transformations to apply, so the two forms are the same on paper. It's still important to remember that there must be the intermediate step between a phonemic representation, which is never spoken, only stored in the lexicon, and a phonetic representation, that which actually comes out of a speaker's mouth.
Let's construct the 3rd person plural of the verb pouvoir, "be able to", given the following mental entry:
"pouvoir" /puvwar/ - root: - stressed (default): /pœ/ - unstressed: /pu/ - plural indicator: /v/
- Retrieve "pouvoir" from the mental lexicon.
- Take /pœ/, the default form.
- Add the plural suffix /v/, giving us /pœv/
- Generate the final phonetic form: [pœv]
/pœ/base + /v/plural + /ɔ̃/1st = /pœv/ + /ɔ̃/stressed -> /puvɔ̃/ ->filter [puvɔ̃]
As you can see, these two apparently irregular verbs can actually be described using exactly the same set of rules!
Now a third example, this time to showcase a surface level phonological "filter": /œ/ in an open syllable (including the end of a word) becomes /ø/.
Using our above method, we make the first person singular form of "vouloir" like this:
- Retrieve "vouloir"
- Take /vœ/,
- Apply surface filters to generate the final phonetic form: [vø]
/vœ/base = /vœ/ ->filter [vø]
This surface filter is the reason that the first 3 forms of the word are [vø], while the last is [vœl]. In [vœl], the syllable is closed, so [œ] doesn't become [ø]. I didn't include this surface filter in the mental entries for the words because this rule applies to all words in the language, not just some verbs. It's on a much higher level than all these alternations, as it's the very last step before actually pronouncing a word. The order goes "lexical entry -> phonemic representation (all the rules we've just examined) -> phonetic representation (after all surface filters)".
So, we've taken this mess:
"vouloir" vø vø vø vulɔ̃ vule vœl
and deconstructed it into a convenient mental model. This approach works for more than just weird verbal forms; it's a way to view seemingly illogical variations in a language and make them make sense, instead of simply memorizing all the forms of a word. It's much more efficient to understand why a word behaves the way it does. Memorizing should always be a last resort.