*ABA – The Goodest Language Universal

Reading time: 5 – 10 minutes

For this month’s blog post, what I’d like to offer is a brief piece about a fascinating universal property of languages, which you may well have never noticed! This apparent language universal is all to do with adjectives, and the way they are built as words.

In English, the majority of adjectives have three forms. For example, the common adjective big can also be bigger and biggest. While all three words share the same sense of ‘bigness’, there is a difference between them as to the degree of bigness. The technical terms for the three are positive, comparative and superlative. The positive (nothing to do with optimism) is the simple form big, which is or is close to the most basic form of the adjective. Bigger is the comparative, because it always implies a comparison with something else. The superlative is biggest, which is used for the thing with the most bigness in a particular context, compared with all others.

Big/bigger/biggest, fast/faster/fastest and thick/thicker/thickest work according to a regular and predictable system, which English has inherited from its Germanic origins: take the positive form, and add -er for the comparative and -est for the superlative. However, not all adjectives follow this system. Sometimes, the endings -er and -est are there, but what they attach to is very different. For instance, it’s not good/gooder/goodest, but good/better/best, as well as bad/worse/worst.

What’s happened here is that words of multiple origins have come together and coalesced into a single grammatical system over time. This phenomenon is known as suppletion. We can find it not only in adjectives; a well-known example is the word went, which started out as a separate verb (namely wend), but has ended up as the past tense of to go. Likewise, the various forms of the verb to be (am, been, was, etc.) come from multiple sources, so we can say that to be is a highly suppletive verb.

Now, when it comes to suppletive adjectives like good, linguists have noticed a very common pattern, or rather patterns, in how they work. In brief, this is the observation:

If the comparative form of an adjective is suppletive, the superlative form must be suppletive too (and vice versa).

This is known as the Comparative-Superlative Generalisation, as formulated by the linguist Jonathan David Bobaljik in his 2012 book Universals in Comparative Morphology.

But what does this mean? Imagine an adjective with a positive form like good. This form we will call A. For historical reasons, its comparative form comes from a different etymological root and looks different to A, like better. This form is suppletive, and we will call it B. According to the generalisation, the third form, the superlative, will never look like A. Either it will share an origin with B, like better and best, or it will have its own separate origin. Either way, it will be suppletive.

So, according to Bobaljik’s generalisation, we find the following patterns:

  • AAA
    • e.g. quick quickerquickest; young younger youngest
  • ABB
    • e.g. good better best; far furtherfurthest
  • ABC
    • e.g. Latin bonusmelioroptimus (‘good’, etc.)

All three patterns fit the generalisation, since we have the regular, non-suppletive pattern AAA and the suppletive patterns ABB and ABC, in which the comparative and superlative forms both show suppletion.

There are even cases where the comparative and superlative forms look different to the positive, but not because of suppletion. For instance, the archaic pattern for the adjective old is old/elder/eldest, all three of which come from the same root. The different vowels are due to a historical sound change, but nonetheless they still follow an ABB pattern.

What are not allowed and not found, as indicated by an asterisk, are the following patterns:

  • *ABA
    • *good bettergoodest
  • *AAB
    • *good gooderbest

Here’s the cool thing: this generalisation, known for short as *ABA, is not only true for English — it seems in fact to be a language universal, something that applies to all languages, everywhere and everywhen.

Whenever someone claims that something is universal, it’s only natural that we want to think of exceptions. What languages do you know? Try it for yourself! Some more examples that come to my (admittedly very Eurocentric) mind include:

  • French: bonmeilleurle meilleur
    • ‘good’, etc. ABB pattern
  • German: vielmehram meisten
    • ‘much/many’, etc. ABB pattern
  • Hungarian: soktöbblegtöbb
    • ‘much/many’, etc. ABB pattern
  • Czech: malýmenšínejmenší
    • ‘small’, etc. ABB pattern
  • Polish: złygorszynajgorszy
    • ‘bad’, etc. ABB pattern
  • Latin: maluspēiorpessimus
    • ‘bad’, etc. ABB pattern

Many exceptions and examples of ABA and AAB patterns have been proposed and argued over, such as on, the Basque word for ‘good’, but even still, the possible exceptions don’t contradict the fact that the Comparative-Superlative Generalisation is remarkably strong.

But why? Why this general ban on *ABA and *AAB? After all, *good/better/goodest would still make sense, thanks to the usual endings. Why can’t it be that only one form of the three is suppletive?

Bobaljik’s explanation for this is the elegant Containment Hypothesis. Simply put:

“The representation of the superlative properly contains that of the comparative”

Bobaljik 2012

What Bobaljik means here is that the superlative is built on the comparative. The comparative and superlative of an adjective are not, despite appearances, equal in their relationship to the positive form. Although younger and youngest look like they are just two possible modifications of young, the Containment Hypothesis states that youngest implicitly contains younger. Therefore, if the comparative form is derived from another origin and is suppletive, it’s no accident that the superlative is too — it has to be!

This idea does make intuitive sense. Unlike the positive form, both the comparative and the superlative involve comparison; bigger implies that something has more bigness than something else, while biggest implies that something has more bigness than everything else!

In some languages, we can see this containment more clearly. Czech is often cited in this regard, as its superlatives, formed with the prefix nej-, visibly build on the comparative.

  • Positive: krásný – ‘beautiful’
  • Comparative: krásnější – ‘more beautiful’
    • [[ krásněj– ] –ší ]
  • Superlative: nejkrásnější – ‘most beautiful’
    • [ nej- [[ krásněj– ] –ší ]]
  • Pattern: AAA
  • Positive: dobrý – ‘good’
  • Comparative: lepší – ‘better’
    • [[ lep– ] –ší ]
  • Superlative: nejlepší – ‘best’
    • [ nej- [[ lep– ] –ší ]]
  • Pattern: ABB

French and other Romance languages form their superlative adjectives by simply adding the definite article to the comparative form. French pire ‘worse’ and plus haut ‘higher’ become the superlatives le pire and le plus haut, literally ‘the worse’ and ‘the more high’.

The Containment Hypothesis is neat, although it does pose further questions. It’s a step forward to recognise that languages build superlatives on comparatives, but still, why is this case? The fact they do does not explain why they do not do it differently. The response to this in turn has been to ascribe *ABA to the controversial concept of Universal Grammar, a deposit of fundamental grammatical principles that are not acquired from one’s linguistic circumstances, but are common to all human brains. It’s not immediately obvious to me why Universal Grammar would care about and include something so specific as the building of superlatives, but it is at least an explanation.

So, that’s *ABA and the Comparative-Superlative Generalisation! It’s a curious phenomenon and a superficially simple language universal hiding right under our collective noses.



Bobaljik, J. (2012). Universals in comparative morphology. MIT Press.

3 thoughts on “*ABA – The Goodest Language Universal

  1. Thank you for another interesting post. Welsh doesn’t not break these rules but it can produce an ABCD suppletive thanks to the equative (as….as) degree. Da (good) has cystal â (as good as), gwell and gorau. Not only are these forms suppletive, but in common with a number of other irregular adjectives, they lack the endings which characterize the different degree of comparison (-ed, -ach, -af).

    Breton also has an equative, but I don’t think that Irish does.

    It would be interesting to hear there are other languages with a distinct equative degree.

    Also are there any languages with more than four degree?

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: