I was talking to Guitarplayer the other day, and he mentioned he wished there was a way to find which features make English weird according the data in the World Atlas of Language Structures. As no simple way is provided, I proceeded to spend five hours examining all the features one by one. Here are the notes I took.
To consider English "weird" within a feature, I used the simple definition "English is not in the largest language group".
A few highlights I personally found interesting:
- English is weird for having no associative plural. Almost half of the languages surveyed use something like "the Marks" to mean "Mark and his people". (Feature 36A.)
- English is weird for having verbal subject agreement but merging all 1st person forms. Some languages have an inclusive/exclusive distinction in the 1st person dual/plural; English runs in the opposite direction by not bothering to distinguish "I do" from "we do"! (Feature 40A.)
- English is weird for its indefinite pronouns derived with generic nouns ("no-body", "every-thing"), something only one quarter (1/4) of the languages surveyed does. More than half of the languages surveyed use something related to interrogatives (so things like "no-who" or "all-what"). (Feature 46A.)
- English is weird for not distinguishing the "with" of accompaniment ("I went with Alice") from instrumental "with" ("I ate with a fork"). More than half of the languages surveyed distinguish these. (Feature 52A.)
- English is weird for having no distributive numerals, namely particular number forms to say things like "one by one" or "two at a time" or "three for each one". Three quarters (3/4) of the languages surveyed have them, formed with a variety of strategies. (Feature 54A.)
- English is very weird for its use of a word ("one") after an adjective to make that adjective stand for an NP, as in "the bad one, the better ones". More than half of the languages surveyed prefer to leave them unmarked (as English sometimes does, e.g. "the rich and the poor"). (Feature 61A.)
- English and various other European languages are very weird for forming perfect TAMs with "have", as in "I have done it". This is simply not found outside Europe in the languages surveyed (some varieties of Mandarin and Minnan Chinese have something like it though, I should add). (Feature 68A.)
- English is weird for putting adjectives before nouns. Over half of the languages surveyed put them after. (Feature 87A.)
- English is weird for its mixed behaviour involving only one negation while using more than one possible construction: "he does nothing" ~ "he doesn't do anything". This is in the context where more than three quarters (3/4) of the languages surveyed prefer the negative agreement of the "he doesn't do nothing" type of construction! (Feature 115A.)
- English and most Germanic languages are very weird for marking polar yes/no questions with word order. It is extremely rare outside the Germanic family. (Feature 116A.)
- It amused me that English used to be a language with "M-T" pronouns ("me ~ thee") but now it's not. (Feature 136A.)
Features where English is weird
Feature 2A. English has a "large" vowel inventory, "large" being defined as having 7 or more vowels. 33% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 3A. English has a low consonant-to-vowel ratio in its phonemic inventory, with not so many consonants compared to the vowels. 10% of the languages surveyed are like this (there are 5 categories in the feature).
Feature 4A. English has a voice distinction in both plosives and fricatives. 28% of the languages surveyed are like this (there are 4 categories in the feature).
Feature 9A. English has a velar nasal phoneme but it can't appear in word-initial position. 19% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 12A. English has a "complex" syllable structure, "complex" being defined as having a syllable structure with more consonants than CCVC. The chapter specifically mentions that the English pattern "is often cited as (C)(C)(C)V(C)(C)(C)(C)". 31% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 15A. English has "right-oriented" stress with stress falling in one of the last three syllables. Note that it really should be one of the last four syllables! 12% of the languages surveyed with weight-sensitive stress are like this (I did not include those reported as having fixed stress in my calculation, as those are basically examined separately in Feature 14A, so there are 7 categories in the feature). (Note, too, that the "12%" mentioned does not include languages with "right-edge" stress on one of the last two syllables; those are counted separately. If we lump them together with the "right-oriented" category English is in, and include all languages surveyed this time, it'd be fair to say, I think, that we find 19% of the languages are like English.)
Comment on Features 14A and 15A: if we lump 14A and 15A together, we find that largest categories of languages are those with fixed penultimate stress or fixed initial stress. English and its stress falling on one of the last four syllables is weird!
Feature 16A. English supposedly has a weight-sensitive stress system based on long vowels + coda consonants. I'm not sure how true this is really, but I admit I don't have a good mental model of how English stress works; I simply use it. 15% of the languages surveyed with weight-sensitive stress are like this (I didn't include those with no weight sensitivity in my calculation, as this feature really just focuses on classifying weight-sensitive stress systems, so there are 6 categories in the feature).
Feature 19A. English has th-sounds. 7% of the languages surveyed are like this (there are 7 categories, 6 of them with a small number of members as this feature is about uncommon consonants (clicks, labial-velars, pharyngeals, th-sounds); most languages, 79%, do not have any uncommon type of consonant).
Feature 22A. English has 2-3 categories per word, which marks it as a language with a particularly small amount of inflectional "synthesis". 17% of the languages surveyed are like this (there are 7 categories in the feature). Because of the many categories in this feature, it's a bit hard to grasp how weird this really is. It may be appropriate to mention that the largest category, 4-5 categories per word, holds 36% of the languages surveyed.
Feature 23A. English supposedly has dependent marking, and 27% of the languages surveyed are supposedly like this. However, this seems to be a mistake on the WALS. According to their own definition of the categories, the nature of English verbs and direct object nouns should be looked at, and neither shows marking for the other. So the real category English should be in is "no marking", and 18% of the languages surveyed are supposedly like this. (There are 5 categories in the feature.) This feature likely needs improvements in its data though, considering English in the wrong category...
Feature 27A. English has no productive reduplication. 15% of the languages surveyed are like this. (There are 3 categories in the feature). It is interesting to note that the bulk of the languages surveyed, 76%, are reported to make use of both full and partial reduplication. It's European and Siberian languages that are strange for having no productive reduplication.
Feature 28A. English has case syncretism in the core cases only. This is probably based in the lack of distinction between subjective "you" and objective "you", whereas the possessive adjectives are distinguished throughout from the core cases (mine, yours, his vs. he/him, hers, ours, theirs). 9% of the language surveyed are like this (there are 4 categories in the feature).
Feature 29A. English has syncretic verbal number/person marking: the plural of "s/he does", "they do", has the same form as "I/you/we do". 30% of the languages surveyed are like this (there are 3 categories in the feature; 41% of languages are not syncretic at all).
Feature 30A. English has three genders (he, she, it). 10% of the languages surveyed are like this (there are 5 categories in the feature; 56% of the languages simply have no gender).
Feature 32A. English has semantic gender agreement. 47% of the languages with a gender system that were surveyed are like this (there are two categories, semantic and semantic-and-formal, when we focus on those with gender).
Feature 36A. English has no associative plurals at all. 16% of the languages surveyed are like this (there are 4 categories in the feature). European, Semitic and many North American languages are the ones that are weird for having no associative plural. Also, it might be worth mentioning the bulk of languages, 44% of them, are reported as having an associative plural identical to the additive plural (so something along the lines of "the Marks" = Mark and his people).
Feature 38A. English has an indefinite article distinct from the number "one". 19% of the language surveyed are like this (there are 5 categories in the feature). Note that this includes languages with no indefinite article; if we focus on languages with an indefinite article, we find 42% of the languages surveyed are like English.
Feature 40A. English has the same form across number in the 1st person: "I do" and "we do". English is very weird for this, including among European languages. 9% of the languages with verbal person marking are like this (there are 4 categories in the feature when we focus on those with person marking).
Feature 43A. English can use its demonstrative pronouns to refer to a person. Unfortunately, the WALS lumps this behaviour with those languages that have demonstratives sharing stems with 3rd person pronouns, so the extent of English's weirdness is particularly unknown here. 23% of the languages surveyed are in the same category as English (there are 6 categories in the feature; the bulk of languages, 44%, have 3rd person pronouns that are totally unrelated to demonstratives).
Feature 46A. English has generic-noun-based indefinite pronouns ("no-body", "no one", "nothing", "some-time", "every-body", etc.). 26% of the languages surveyed are like this (there are 5 categories in the feature; the bulk is taken by interrogative-based indefinite pronouns in 60% of the languages).
Feature 50A. English has "additive-quantitively asymmetrical" case marking. "Asymmetrical" case marking means there are NP categories that have less cases than others. "Additive-quantitively asymmetrical" case marking means that most NPs in a language are marked for less cases than a certain select type of NPs, as in English "Oscar"/"Oscar's" vs. "he/him/his" (three cases are only found in personal pronouns and in "who/whom/whose"). 29% of the languages with case marking that were surveyed are like this (5 categories in the feature if we focus on languages with case marking). The bulk, 44%, are symmetrical.
Feature 51A. English has no case affixes or case-marking adpositional clitics. 37% of the many (1031) languages surveyed are like this (there are 9 categories in the feature). It is worth mentioning the bulk of languages, 44%, uses case suffixes only. (Postpositional clitics account for 12% of the languages too.)
Feature 52A. English does not distinguish accompaniment "with" ("I went with Ted") from instrumental "with" ("I did with a hammer"). 24% of the languages surveyed are like this (there are 3 categories in the feature; 66% of the languages do clearly make the distinction).
Feature 53A. English has suppletive ordinal numbers for "1st" (first) and "2nd" (second). 19% of the languages surveyed are like this (there are 8 categories in the feature; the bulk, 34%, consists of languages with suppletive "1st" (first) but a normal "2nd" derived with an affix (two-th)).
Feature 54A. English has no distributive numerals. 25% of the languages surveyed are like this (there are 7 categories in the feature, accounting for different types of distributive numeral formation).
Feature 56A. English has formally similar "conjunctions" and universal quantifiers that don't involve interrogative morphemes. "Conjunctions" here are broadly defined to mean any sort of phrase connector, including more adverbial ones like "also" and "only". English counts apparently because of "all" and "also" (all-so). 28% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 57A. English has no possessive affixes. 29% of the languages surveyed are like this (there are 4 categories in the feature, the bulk taken by the 39% that has possessive suffixes, mostly in Eurasia/Africa/Oceania).
Feature 61A. English marks adjectives without nouns using a word placed after the adjective ("one" as in "the good ones"). 6% of the languages surveyed are like this (there are 7 categories in the feature). Most languages, 59% simply do not mark adjectives that constitute a whole NP, as English sometimes does e.g. "the rich and the poor", "the well-intentioned and the mean-intentioned", "the Chinese and the French".
Feature 62A. English has a "mixed" strategy for the arguments of action nominal constructions. As the WALS explains, English uses a variety strategies, namely "the Possessive-Accusative type (Peter’s singing the Marseillaise), [...] the Ergative-Possessive type (the destruction of the city by the enemy) and [...] the Double-Possessive type (the enemy’s destruction of the city)". 8% of the languages surveyed are like this (there are 8 categories in the feature).
Feature 68A. English forms its perfect TAMs with a possessive ("I have done it", "you had done it", "we will have done it"). 3% of the languages surveyed are like this (there are 4 categories in the feature). This is one of the most remarkably weird characteristics of European languages guys!
Feature 70A. English has no morphologically distinctive imperatives. This is actually very weird for a European language, the only other European language surveyed that's like this, out of the many that are included, is Hungarian. 22% of the languages surveyed are like this, mostly found among the highly analytic languages of Southeast Asia (there are 5 categories in the feature).
Feature 71A. English forms if its prohibitives with the normal negative and the normal imperative: "you don't do it", "do it!" ~ "don't do it!". 23% of the languages surveyed are like this (there are 4 categories in the feature). It might be interesting to note the bulk is found in the 37% of the languages surveyed that use the normal imperative with a special negative.
Feature 75A. English uses verbal constructions ("may", "must") for epistemic modality ("he may have arrived"), like most European languages. 27% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 76A. English has overlap in the expression of situational and epistemic modality, as the epistemic auxiliaries "may" and "must" can have the so-called "situational" meanings "be allowed to, be able to" and "be obliged to". This is common in Europe, but a bit less common elsewhere. 17% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 78A. English doesn't morphologically code evidentiality. 43% of the languages surveyed are like this (there are 6 categories in the feature, accounting for different types of evidentiality coding).
Feature 79A. English has suppletion according to tense ("I am ~ was", "I go ~ went", less strongly "I buy ~ bought"). 19% of the languages surveyed are like this (there are 4 categories in the feature, most lanugages, 64%, do not have suppletion for tense or aspect).
Feature 81A. English is SVO. 35% of the languages surveyed are like this (there are 7 categories in the feature). As we all know, there are more SOV languages in the world; they consitute 41%.
Feature 83A. English is VO. 46% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 85A. English has prepositions. 43% of the languages surveyed are like this (there are 5 categories in the feature). Half of the languages, 49%, use postpositions.
Feature 86A. English has no dominant order in its genitives, sometimes using "[noun]'s [noun]" and sometimes using "[noun] of [noun]". 8% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 87A. English puts adjectives before nouns. 27% of the languages surveyed are like this (there are 4 categories in the feature). Most languages, 64%, put the adjective after the noun.
Feature 88A. English puts demonstratives before nouns. 44% of the languages surveyed are like this (there are 6 categories in the feature). I should point out most other languages, 46% of the total, simply put a demonstrative word after the noun. It's the ones that use demonstrative affixes that are weird, like a whole bunch of the Nilo-Saharan languages.
Feature 89A. English puts numbers before nouns. 42% of the languages surveyed are like this (there are 4 categories in the feature). Europe and the northern half of Asia put numbers before nouns universally, but much of the rest of the world does otherwise, so that 53% of the total puts numbers after nouns.
Feature 93A. English has wh-movement in content questions. 29% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 95A. English is VO and has prepositions. 40% of the languages surveyed are like this (there are 5 categories in the feature). The other huge group of languages is being OV and having postpositions, taking 41% of the total. There are languages that are VO with postpositions and OV with prepositions but they are quite weird (4% and 1% of the total respectively).
Feature 97A. English is VO and puts adjectives before nouns. 9% of the languages surveyed are like this (there are 5 categories in the feature).
Feature 99A. English has nominative-accusative alignment in its pronouns. 35% of the few languages surveyed are like this (there are 7 categories in the feature). I should note this feature doesn't have many languages surveyed. With more languages, the proportion could easily change.
Feature 101A. English has obligatory pronouns in subject position to express pronominal subjects. 12% of the languages surveyed are like this (there are 6 categories in the feature). Most languages, 62%, prefer to use verbal affixes.
Feature 102A. English only marks subjects in its verbs. 19% of the languages surveyed are like this (there are 5 categories in the feature).
Feature 105A. English has a mixed strategy for argument marking in the verb "to give". It is possible to say either "I give you a book" or "I give a book to you". 11% of the languages surveyed are like this (there are 4 categories in the feature).
Feature 107A. English has passive constructions. 43% of the languages surveyed are like this (there are 2 categories in the feature). All of Europe has passive constructions, but most other areas have a mixture of languages with and without passive constructions.
Feature 110A. English has a sequential periphrastic causative construction ("they make me work"), but not a purposive one. "Purposive" is defined as having some sort of subjunctive or allative marker that suggest purpose (*"they made me to work", *"they forced me so that I work"), as opposed to strict sequences of events. 30% of the languages surveyed are like this (there are 3 categories in the feature). Rather few languages are surveyed, but the WALS explains this is because periphrastic causatives are often not discussed in reference materials...
Feature 114A. English has asymmetric standard negation with regard to emphasis, as the difference between positive "he works" and "he does work" gets lost in the negated construction "he does not work". Many such strange types of asymmetrical negation can be found in grammars, and for the sake of simplifying things, the WALS simply puts most of them in a miscellaneous category: "asymmetric in other grammatical categories". About half of the languages with asymmetric negation belong to the miscellaneous category (there are 6 categories of asymmetric negation distinguished in the feature).
Feature 115A. English shows mixed behaviour in negation involving indefinite pronouns: "they do nothing" ~ "they don't do anything". 6% of the languages surveyed are like this (there are 4 categories in the feature). It is interesting to note that the vast majority of languages, 83%, make use of verbal negative agreement of the "they don't do nothin'!" type.
Feature 116A. English yes/no questions are marked with a particular word order: "she can bake" ~ "can she bake?". 1% of the languages surveyed are like this (there are 7 categories in the feature). This is a common thing among Germanic languages, but it is extremely weird at the global level. Outside Europe one may find this in Manggarai (Central Malayo-Polynesian), Palauan (a rather weird Malayo-Polynesian language admittedly), Hup (a language from a tiny family in northwestern Brazil) and Warekena (Arawakan; northwestern Brazil). Most languages, 61%, use a question particle.
Feature 118A. Predicative adjectives are not encoded as verbs (we don't have to say "it greens" for "it is green"). 34% of the languages surveyed are like this (there are 3 categories in the feature). 39% of the total encode it as verbs and the rest used a mixed strategy. Europe and the northern half of Asia universally agree with English, but not much of the rest of the world.
Feature 119A. English uses the same verb, "to be", for both nominal predication ("she is tall", "she is an engineer") and locational predication ("she is in New York"). 30% of the languages surveyed are like this (there are 2 categories in the feature, "different" and "identical").
Feature 121A. English uses a particle in its comparative constructions, "than", like most European languages. 13% of the languages surveyed are like this (there are 4 categories in the feature).
Feature 122A. English uses case-marked relative pronouns to mark relative clauses. 7% of the languages surveyed are like this (there are 4 categories in the feature).
Feature 123A. English uses a relative pronoun to relativize obliques ("the knife that I cut it with", "the woman that gave me the gift"). 12% of the languages surveyed are like this (there are 5 categories in the feature).
Feature 125A. English uses both "balanced" and "deranked" purpose clauses, which means that its purpose clauses both look and don't look like independent clauses. We can say "I bought it for you to use it" or "I bought it so that you use it". 18% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 126A. English uses both "balanced" and "deranked" when-clauses, which means that its when-clauses both look and don't look like independent clauses. We can say "when beginning to swim, make sure to relax" or "when you begin to swim, make sure you relax". 22% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 127A. English uses both "balanced" and "deranked" reason clauses, which means that its reason clauses both look and don't look like independent clauses. We can say "punish him for saying that" or "punish him because he said that". 22% of the languages surveyed are like this (there are 3 categories in the feature).
Feature 132A. English has a large number of non-derived basic colours (six). 24% of the languages surveyed are like this (there are 7 categories in the feature).
Feature 133A. English has a large number of basic colours (eleven). 9% of the languages surveyed are like this (there are 7 categories in the feature).
Feature 134A. English distinguishes green and blue. 25% of the languages surveyed are like this (there are 7 categories in the feature).
Feature 136A. English doesn't have M-T pronouns, or at least not anymore ("me", "thee"). It's weird for a "Nostratic" language at least (there are 3 categories in the feature).
Feature 138A. English has its "tea" word derived from Min Nan Chinese te. 37% of the languages surveyed are like this (there are 3 categories in the feature).
Features where English is like most languages, or in any case like the largest bulk of languages
1A, 5A, 6A, 7A, 8A, 10A, 11A, 13A, 14A, 17A, 18A, 20A, 21A, 21B, 24A, 25B, 26A, 33A, 34A, 35A, 37A, 39A, 41A, 42A, 45A, 47A, 48A, 55A, 58A, 58B, 59A, 60A, 63A, 64A, 65A, 66A, 67A, 69A, 72A, 73A, 74A, 77A, 79B, 80A, 82A, 84A, 90A, 90C, 91A, 92A, 94A, 96A, 98A, 100A, 103A, 104A, 106A, 108A, 108B, 109A, 109B, 111A, 112A, 113A, 117A, 120A, 124A, 128A, 129A, 130A, 131A, 135A, 137A, 137B, 142A, 143A, 143E, 143F, 143G, 144B, 144D, 144H, 144I, 144J, 144K.
Features where English seems to be unlike most languages, but not really
Feature 25A assumes the Feature 23A category English is in is correct, but it is not. So while English appears to be weird in Feature 25A, it is not once you correct the classification.
Feature 31A focuses on languages with gender systems, to see whether they're sex-based or non-sex-based. 75% of the languages with gender systems that were surveyed have sex-based systems like English does, so English is normal as far as sex-base-ness in gender systems goes.
Feature 44A shows English distinguishes gender in 3rd person singular personal pronouns only. This is the case for 49% of the languages with gendered personal pronouns that were surveyed, the largest category of the bunch. English is normal when it comes to gendered pronouns here, although most languages, 67%, do not have gender distinctions in personal pronouns.
Feature 49A. English nouns have two morphological cases, presumably direct and genitive, e.g. "Julia" and "Julia's". Having morphological case as defined by WALS is not weird in itself, 53% of the languages surveyed do. And if we focus on languages with morphological case for sure (7 categories in the feature, based on the number of cases), we find 17% of the languages surveyed have 2 cases like English does. The category for 6-7 cases has more languages, but it combines languages with 6 cases and 7 cases. If you're going to have case, the 2-case category seems to be what is most common.
Feature 144A. English has the word order subject-negative word-verb-object. I am only mentioning this feature because the largest categories in this feature are the category of languages without negative words at all (not the focus of the feature to begin with), and the miscellaneous bag category.
