Let's make a summary of the main language families

[Squall]

In order to make conlangs following rules of some families, we need to know the family first.

Indo-European: The original language has few fricatives and lots of plosives. The nouns have three genders and multiple cases. It has dual along with plural, which are part of the declensions. The verbs have huge conjugation tables with many tenses, aspects and voices. It has conjugation classes and declension classes, which are defined by a thematic vowel.

Slavic (Russian, Czech): It is a typical Indo-European sub-family, because it keeps genders, cases and agreements. It has an average number of vowels, lots of fricatives, lots of palatalization and lots of consonant clusters. Rather than definite articles, it uses word order and cases to inform definiteness.

Germanic (Danish, Dutch): It has a large number of vowels and it has diphthongs that may have composing monophthongs absent in the language. It has small words and complex syllables. It is stress-timed. Unvoiced plosives are aspirated. It is also vulnerable to vowel shifts. It lost many Indo-European features: cases and conjugations. Genders were lost or simplified in most languages. It developed vowel mutation, which gave birth to many irregular verbs and irregular plurals. It uses auxiliary verbs and modal verbs. Its proto-language uses the verb-second word order, but only German and Dutch keeps it.

Romance (Castilian, Italian): Romance has an average number of vowels, simple syllables and many multisyllabic words. It has two genders with agreements and no cases. It uses both conjugation tables and auxiliary verbs. It has prepositions and complex rules regarding oblique pronouns. The word order is SVO. There are two black sheeps, Romanian and French. Romanian is not part of the sprachbund and has Slavic influence. French is phonologically very different; it has Germanic influence, but it does not sound like Romance nor Germanic. French, however, has grammatical features of the European sprachbund.

Semitic (Hebrew, Arabian): It is an Afro-Asiatic branch. This is the family of the tri-consonantal roots. It has two genders as well. It has few vowels and lots of consonants. It has many guttural consonants (but Maltese is an exception).

Sinitic (Mandarin, Cantonese): It is a Sino-tibetan branch. It has a complex tone system and the syllables are complex. Most roots are monosyllabic. It is very isolating, but has a few affixes.

Bantu (Swahili, Shona): It has many word classes and their corresponding concords. Plurality and definiteness are informed in the concords. The syllables are simple, usually CV. It has tones, but its features are very different from the Asian systems. It has nasal syllables and nasal vowels. (Swahili is the exception, because it lacks tones and nasal vowels.)

Uralic (Finnish, Hungarian): It is based on agglutination with many suffixes. It has many cases. The grammar makes it difficult to insert loanwords. It has vowel harmony and many vowels.

Malayo-Polynesian (Malay, Tagalog): It is an Austronesian branch. It has a small to average number of vowels and few fricatives. Verbs have affixes, but tense, negation and aspect are not informed in the verb. It uses adverbials to inform if the action "has occurred", "is occurring" or "has not occurred yet". The plural is informed by reduplication of the noun. It is verb-initial (the exception is Indonesian languages).

There are many scattered individual languages that are part of no families. An example is Basque.

Notable sprachbunds

Japanese and Korean (Japanese, Korean): The most notable feature is the politeness in the grammar. It is topic-prominent and strongly head-final. The sentences ends with the verb. It has lots of grammatical particles (postpositions). Phonologically, Japanese are Korean unrelated.

Altaic (Turkish, Mongolian): It isn't an established language family, but has many similarities. It is similar to Uralic in many aspects, but they are not part of the same family.

[Frislander]

I can already think of an exception to your 'rules' in Sino-Tibetan, as multiple branches of the family (such as Tibetic) are at least mildly agglutinating and the rGyalrong languages are even polysynthetic. Even Mandarin has a small set of affixes.

[Frislander]

There are many families in North America. Some languages, such as Navajo, are so unusual that they were used for cryptography during the WWII. They are strongly verb-centered and work very differently from most languages.

In addition, I also have problems with this paragraph. Firstly, there was nothing especially about Navajo (or Choctaw or Cherokee or any of the other languages used by the US Army for military purposes) that leant them especially for this purpose (c.f. the use of Welsh by the Royal Welsh Fusiliers for similar purposes): I doubt the commanders had much idea at all about the actual structure of said languages.

There is also very little about Navajo specificaly that makes it ideal for this purpose over, say, Mohawk or Lakota: the real factor in deciding what language(s) would be used was for the most part just the size of speakerbase.

Secondly, the idea that they work 'very differently' from most languages is misleading. While it is certainly true that Native American languages are in many ways distinctive, that does not mean to say that they are somehow radically different from other languages of the world. For instance perhaps the defining feature of most Native American languages (though certainly not all), polysythesis, is also found in South America, Northern Australia, New Guinea, the Caucasus, eastern Siberia EDIT: and northern India. There is little to suppose that there is something about said feature that confines it to North America.

[GrandPiano]

I can already think of an exception to your 'rules' in Sino-Tibetan, as multiple branches of the family (such as Tibetic) are at least mildly agglutinating and the rGyalrong languages are even polysynthetic. Even Mandarin has a small set of affixes.

"The roots are monosyllabic" is also not entirely true. Mandarin has several disyllabic roots that can't be broken up into monosyllabic morphemes, e.g. 蝴蝶 húdié "butterfly" and 葡萄 pútao "grape".

[Avo]

These "summaries" have so many problems that I don't even know where to start.

/edit: I was in a hurry earlier and felt bad for just leaving this one-liner, so I wanted to get back to it. But I literally don't know where to start. Your summaries ("Language family: Features X, Y and Z" hardly qualifies as a summary in my opinion) are full of inconsistencies, nonsense sentences and things that are simply wrong. What does China have to do with a language having tone? Since when is Korean a Japonic language and how is politeness _the_ main feature of Japanese grammar? What do Japanese and Korean even have to do with Germanic and Romance languages? What is "Uralic grammar makes it difficult to insert loan words" even supposed to mean, especially since both languages you mentioned have had a massive influx of loanwords from several different sources?
Altaic isn't even an established language family and it sure isn't "like Uralic". French is a Romance language, thus I'd say it pretty much sounds like one. Its phonology doesn't even stand out among its surrounding related languages. And how can you even lump massive and insanely diverse language families like Sino-Tibetan or Austronesian into three short sentences? Or the entirety of language families in North America? Navajo isn't "so unusual" and it wasn't used in cryptography during the WWII because it's so unlike every other language on earth, it just happened to be one of the more vital Native American languages that hadn't been extensively documented by linguists yet at the time, especially not by Germans. What are the languages in parentheses supposed to be, example languages? Your choices seem pretty random and meaningless to be honest.

[Xing]

Semitic (Hebrew, Arabian): This is the family of the tri-consonantal roots. It has two genders as well. It has few vowels and lots of consonants. It has many guttural consonants (but Maltese is an exception).

Semitic languages are no language family; they are a sub-branch of the Afro-Asiatic family.

[Iyionaku]

It lost many Indo-European features: genders, cases and conjugations. (...) It uses the verb-second word order.

The first sentence might appear for your example languages (although all Germanic languages have gender, the North Germanic ones only have a differing), but German, Faroese and Icelandic have all that pretty much. Opposed to that, Verb-second order is used almost exclusively on continental western Europe (German and Dutch dialects).

[KaiTheHomoSapien]

Lol you guys are really tearing this guy apart, but I think it's a cool topic to discuss, the OP just shouldn't have presented the original post as cold hard fact.

A lot of language families have certain characteristics that all their member languages share, but sometimes the families are really diverse within themselves.

Here are some language families that I've always been interested in, but don't know a whole lot about (so nothing I say here is meant to be definitive fact!):

Northwest Caucasian (Abkhazo-Adyghean) - Circassian and Abkhaz are members. Simple vowel systems, but elaborate consonant systems.
Northeast Caucasian (Nakho-Dagestanian) - Avar, Lak, Chechen, and Ingush are members. This one also has a lot of elaborate consonant systems, as well as interesting gender systems. Might be connected to Hurro-Urartian in an "Alarodian" family.
South Caucasian (Kartvelian) - Georgian is the main member and the Caucasian language with the most speakers by far. Languages are mainly ergative and also have a lot of consonants (some people try to connect Basque with Kartvelian). I think all the Caucasian language families influenced each other even if they're not genetically related. I think it's unlikely they're connected to Indo-European too, but I guess it's possible.

Cariban - Hixkaryana, the "OVS" language, is the only member I know. This one is spoken in northern South America. I'd like to research this family more.
Great Andamanese - extinct family on the Andaman islands. I've always been interested in this one because we know so little about it and because the Andaman Islands are so interesting--they include an island that's been largely untouched by modern civilization.

[Trebor]

Lol you guys are really tearing this guy apart, but I think it's a cool topic to discuss, the OP just shouldn't have presented the original post as cold hard fact.

Well said.

Of Semitic, I'll add that Amharic, Tigrinya, and Tigre are members, but don't have vowel systems like that of Arabic. Tigrinya and Tigre have velar and pharyngeal fricatives, while Amharic has neither. All three lack the "emphatic" consonants of Arabic, but make up for it with ejective plosives and fricatives/affricates. Also, standard modern Hebrew has neither pharyngeal nor "emphatic" consonants.

Of Bantu, I'll clarify that its members tend to possess a large number of noun classes, comparable to masculine, feminine, and neuter in Indo-European but not gender-based. Most noun classes come in pairs, with one prefix on a root marking singular and another marking plural. Some roots have a plural that would not be expected from its singular form, and nouns in one class can move to another and thereby take on an augmentative, diminutive, abstract, and even (Finnish case-style) locative sense. Some Bantu languages use consonant+vowel (CV) noun-class prefixes, while others add a vowel (V) pre-prefix, which can be used to convey information about definiteness.

Additionally, phonotactics are generally restrictive, permitting vowel (V) and consonant+vowel (CV) syllables, with the caveat being that what could be analyzed as nasal+obstruent and consonant+semivowel clusters are also very common. (Plus, there are languages in and around Cameroon and the DR Congo that allow syllables to end in a consonant.) Tone is indeed almost ubiquitous, but it looks very different from the feature as it appears in southeast Asia. Most languages have only high and low tones on short vowels (exceptions being Sukuma and Nzadi), with rising and falling possible on long vowels. Tone can distinguish the lexical meaning of words, but also their morphological/syntactic function. On the subject of Bantu phonology, I recommend this PDF.

As well, it's important for conlangers to bear in mind that Africa has so much more to inspire them than just Bantu. Languages worth investigating include Yoruba, Igbo, Fongbe/Ewe, Dagbani, Mooré/Mossi, Bambara, Fula, and Wolof (tentatively or indisputably belonging to the larger Niger-Congo family); Dholuo, Dinka, Nuer, Ngambay, and Ingessana/Gaahmg (possibly or definitely Nilo-Saharan); Gurage languages like Chaha (related to Amharic); Oromo and Somali (from the Cushitic branch of Afro-Asiatic, related to Semitic); and Hausa (from the Chadic branch). There's also the Ubangian family/Niger-Congo subgroup, such as Sango, but I haven't found much about it online.

Edit: Minor revisions.

[Squall]

Thank you for the correction. For that reason I posted this list in the Beginners section rather than the Linguistics section.

I have fixed most summaries.

Semitic languages are no language family; they are a sub-branch of the Afro-Asiatic family.

Because it is difficult to generalize a primary family, I am using sub-branches in some parts.
So I had to change into Sino-Tibetan to Sinitic, because of over-generalization.

What do Japanese and Korean even have to do with Germanic and Romance languages?

It is a comparison. Germanic and Romance are phonologically very different, but they came from the same mother language.
Japanese and Korean are phonologically very different as well. I do not know the name of the source language, but it is not Altaic.

What is "Uralic grammar makes it difficult to insert loan words" even supposed to mean, especially since both languages you mentioned have had a massive influx of loanwords from several different sources?

There is an article about Finnish that says that it has few cognates with European languages and it prefers creating words with existing roots rather than borrowing.

French is a Romance language, thus I'd say it pretty much sounds like one. Its phonology doesn't even stand out among its surrounding related languages.

It does not sound like other Romance ones. It underwent vowel shift that other Romance did not making the pronunciation unrecognizable. It has round front-vowels and doesn't have stress.

And how can you even lump massive and insanely diverse language families like Sino-Tibetan or Austronesian into three short sentences?

So I have to divide into branches (Sinitic and Malayo-Polynesian) or describe the proto-language.

What are the languages in parentheses supposed to be, example languages? Your choices seem pretty random and meaningless to be honest.

Yes.

Of Bantu, ...

It is a really good summary.

As well, it's important for conlangers to bear in mind that Africa has so much more to inspire them than just Bantu.

There is a family of clicks, but I am not familiar with it.

[KaiTheHomoSapien]

The "family of clicks" you're thinking of is probably Khoi-San, but it was later determined to not really be one family, but rather a collection of smaller families that all contained languages with clicks. But people still say Khoi-San to refer to the click languages in general.

[Iyionaku]

What do Japanese and Korean even have to do with Germanic and Romance languages?

It is a comparison. Germanic and Romance are phonologically very different, but they came from the same mother language.
Japanese and Korean are phonologically very different as well. I do not know the name of the source language, but it is not Altaic.

Japanese and Korean are (by state of research) not related to each other. Japanese forms its own language family with the few Ryukyu languages. Korean counts as a language isolate, although some scolars want to put it together with Turk languages, Mongolian languages and Tungusian languages to the altaic language family.

However, Japanese and Korean have severe simliarities in their language typology that might have evolved in a sprachbund.

(Source: Wikipedia)

The "family of clicks" you're thinking of is probably Khoi-San, but it was later determined to not really be one family, but rather a collection of smaller families that all contained languages with clicks. But people still say Khoi-San to refer to the click languages in general.

Khoi-San are generally all African languages that are neither Afroasiatic, nor Nilo-Saharan, nor Niger-Kongo.

[qwed117]

What do Japanese and Korean even have to do with Germanic and Romance languages?

It is a comparison. Germanic and Romance are phonologically very different, but they came from the same mother language.
Japanese and Korean are phonologically very different as well. I do not know the name of the source language, but it is not Altaic.

Japanese and Korean are (by state of research) not related to each other. Japanese forms its own language family with the few Ryukyu languages. Korean counts as a language isolate, although some scolars want to put it together with Turk languages, Mongolian languages and Tungusian languages to the altaic language family.

However, Japanese and Korean have severe simliarities in their language typology that might have evolved in a sprachbund.

(Source: Wikipedia)

The "family of clicks" you're thinking of is probably Khoi-San, but it was later determined to not really be one family, but rather a collection of smaller families that all contained languages with clicks. But people still say Khoi-San to refer to the click languages in general.

Khoi-San are generally all African languages that are neither Afroasiatic, nor Nilo-Saharan, nor Niger-Kongo.

Not quite. Nilo-Saharan is generally all clickless African language that isn't Afro-Asiatic nor Niger-Congo and Khoi-San is all click languages spoken south of the equator. There are some exceptions though, like Laal, and Hadza that don't really go anywhere

[KaiTheHomoSapien]

Yeah, Hadza seems to be an isolate with clicks. There's also Bangime, an isolate spoken in Mali that is clickless and not grouped with Nilo-Saharan. Nilo-Saharan is pretty tentative as it is, though.

[qwed117]

Yeah, Hadza seems to be an isolate with clicks. There's also Bangime, an isolate spoken in Mali that is clickless and not grouped with Nilo-Saharan. Nilo-Saharan is pretty tentative as it is, though.

I'll be brutally honest: The reason Bangime is not grouped with Nilo-Saharan is that Greenberg died before it got reclassified into an isolate.

[gach]

I'll be brutally honest: The reason Bangime is not grouped with Nilo-Saharan is that Greenberg died before it got reclassified into an isolate.

Makes you confident about the received picture of language classification, doesn't it.

[KaiTheHomoSapien]

Any of those languages has the potential to be classified just as well as IE languages, but it's hard when the language is spoken by the people of a small hard-to-access village in a very poor nation. But the fact that we're able to study and classify IE languages in such a detailed way shows it can be done for any family. It's just a matter of the difficult field work (and the unwritten nature of many of these languages).

[WeepingElf]

The African languages classification algorithm goes like this:

1. If it shows signs of being related to Semitic, it goes to Afroasiatic.
2. If it shows signs of being related to Bantu, it goes to Niger-Congo.
3. Otherwise, if it has click consonants, it goes to Khoisan.
4. Otherwise, if it doesn't have click consonants, it goes to Nilo-Saharan.

[Creyeditor]

The African languages classification algorithm goes like this:

1. If it shows signs of being related to Semitic, it goes to Afroasiatic.
2. If it shows signs of being related to Bantu, it goes to Niger-Congo.
3. Otherwise, if it has click consonants, it goes to Khoisan.
4. Otherwise, if it doesn't have click consonants, it goes to Nilo-Saharan.

Yeah, and I think that's really the definition of Khoisan, there's nobody claiming anymore it's really one language family.

[Trebor]

If Nilo-Saharan, Khoisan, etc. are disproven/remain uncertain, surely we (in the general sense) can at least talk about and summarize those subgroups which are clearly legitimate. What will benefit conlangers reading this thread is information about how less-known languages operate, not a complete picture of how those languages may or may not be related to one another.

So, who's up for tackling, say, Eastern (e.g., Dinka) vs. Central (e.g., Ngambay) Sudanic?


Edit: It would be very helpful for a work like "Ngiti: A Central-Sudanic Language of Zaire" to become available online*. In the meantime, I can link to "The "Macro-Sudan belt": Towards identifying a linguistic area in northern Sub-Saharan Africa".

[More edits: And "A Brief Grammatical Sketch of Ngambay" along with "A grammar of Gaahmg, A Nilo-Saharan Language of Sudan".]

*I highly doubt OCR software can handle the text, or I'd look into borrowing a hard copy to give an overview of Ngiti for this thread.