The real answer is that all European languages used to be this way, and western Europe at some point* started relying more heavily on fixed word orders and prepositions instead of suffices, but further east in Europe they did not have this trend.

* in romance languages my understanding is that the big deal was that consonants at the end of words and sometimes syllables were no longer pronounced, therefore it was impossible to distinguish a bunch of suffices. eg. final -m nasalized the vowel before it and then later was dropped entirely. Or you could no longer distinguish between -us and -um because they were both -u (later -o).

The creole languages I'm familiar with tend to be at least as analytic as the most analytic of its parent languages, and possibly more analytic than any of them.

Haitian creole is the most analytic language I'm familiar with. I gather it's about on par with languages like Ewe, but I know next to nothing about Ewe, so I'm not sure on that. I don't think it absorbed a single inflection from French, though.

Modern English, arguably descended from a creole of Old English and Norman French, is less inflected as either of those parent languages. For example, both Old English and French have grammatical gender, but middle and modern English retain it only for third person singular pronouns.

I don't know much about the history of the development of the Romance languages, but I had always guessed that they are at least somewhat a result of creolization of Latin with local language groups, and, if my speculation (it's not even a hypothesis) is at all correct, that would imply that they all would tend to lose inflection, including agglutination, in the process.

So, all that said: Perhaps the reason eastern European languages are more like Latin in this respect is, ironically, because they were less influenced by Latin?

Well, there was something called "Vulgar Latin" which varied by region, and I do think the biggest change was that differing pronunciations forced grammar changes as you could no longer differentiate with "silent" letters.

The loss of inflection due to final consonants no longer being pronounced explains quite a bit of things I found mysterious when learning romance languages without much Latin knowledge.

Some examples using Spanish:

* Spanish lost most neuter forms, but why does it distinguish este and esto [both meaning "this"]? Well, in Latin it was iste and istud respectively, so dropping final consonants still preserves distinct forms: iste and istu. Contrast that with most masculine/neuter contrasts: bonus and bonum both become bonu. (Hence, only one bueno in Spanish.)

* When I learned the language I was always confused why imperfect or present subjunctive first person singular is indistinguishable from third person. (eg. yo amaba, el amaba, I loved, he loved) Well, in Latin, those differed by a final consonant: ego amabam vs. amabat for third person.

But in Spanish, unlike French for example, all letters are still pronounced. So I'm not sure this reasonning holds.

I think we are talking about differing times in evolving languages. All the letters of Latin are most certainly not pronounced in Spanish. That is why a bunch of words end in -o instead of -us or why no verb conjugation ends in -m or -t anymore.

At some point people in Spain stopped writing Latin and started writing Spanish, at which point spelling matched pronunciation. But that was not today's Spanish. There were big pronunciation and spelling changes around the 15th century IIRC - see this article for some hints of how it used to look and sound. https://en.wikipedia.org/wiki/Old_Spanish_language

The current "spelling matches pronunciation" thing only occurred because the spelling was updated when pronunciation changed. It is kind of a time capsule, needs to be periodically "maintained", not innate to the language.

Also important to note that there are places where not every letter is pronounced in modern Spanish. Eg. Many accents aspirate or omit S in some positions. Some people omit D between vowels.

A few more examples:

1) z and s, or ce and se, are pronounced the same everywhere except some parts of Spain. I guess not an example of silent letters but still, an example of a distinction that existed historically/etymologically but that no longer does in the spoken language for the vast majority of speakers.

2) The h is silent in every dialect despite having usually represented a real sound in Latin

> z and s, or ce and se, are pronounced the same everywhere except some parts of Spain.

I didn't cite that because I'm already blabbing too much about this topic, and I didn't want to overwhelm people, but yes. Also, as cited by the sibling comment, the merger of /b/ and /v/. Also, in many accents, but not all, Y and LL are the same. Also, consonants like /b/ or /g/ are sometimes swallowed when followed by /w/.

Also <b> and <v> have merged in a lot of versions of Spanish. The <v> is postclassical but represents /w/. A minimal pair for these in Latin is bis 'twice' and vis 'force'. (Well, not quite, because of different vowel lengths... let's try bilis 'bile' and vilis 'vile'.)

So, I can't speak to Spanish, but I can speak to French. There, the suffixes that are silent in Metropolitan French are still pronounced in some regional dialects, but the grammar doesn't really change (that I've noticed).

That leads me to think that it could just as easily be that grammar changes render the agglutination superfluous, and it subsequently withers away like a vestigial organ.

You're not quite right. Classical Latin had a lot of things that virtually no modern language has: a fully developed passive conjugation, a fully developed conjunctive, and - the cherry on the cake - a rigorously expressed separation between phases (consecutio temporum). Latin is quite magical, and even if it was actually more crude than some of its contemporaries like Ancient Greek and Sanskrit, it's still sweeter than virtually all modern languages, some of which don't even have a past perfect tense or more than one conjunctive form.

I have no idea what "fully developed passive conjugation" is, but my language is very old and I imagine very difficult for English speaker to learn. There is an example how fun it can be: http://www.lituanus.org/1987/87_1_04.htm

It means that every conjugation of an active verb has a corresponding passive form, as in your link: esąs - buvęs - būsįas versus esamas - būtas - būsimas . But that's underselling it a bit, because Latin actually had a full 3-dimensional matrix of conjugations: time-based tense (past, present, future), goal-based tense (completed or ongoing), and action (active or passive), with accompanying participles for each case. To compare with your link, I notice Lithuanian seems to lack passive participles ("while being beaten" is the corresponding example), while keeping some of the circumstantial conjugations (frequency, intent, necessity, conjunct phrases) that older languages did have, but Latin did away with.

In more analytical languages, most of these constructs have been replaced with auxiliary verbs (as in English: I see versus I'm seen, I'm watching versus I'm being watched). With full conjugations, entire sentences can change in tense or action without affecting word order. This may not seem that special if your native language is like that, but as the OP started out saying, it allows for much more expressiveness in (predominantly written) language.

If I got it right, Lithuanian does have passive participle (neveikiamasis dalyvis). In the case of "while being eaten" it would be "valgomas". "While being beaten" -> "mušamas". He was silent while being beaten -> Mušamas jis tylėjo.

To quote Wikipedia:

    the Lithuanian language is unique for having fourteen different
    participial forms of the verb, which can be grouped into five
    when accounting for inflection by tense. Some of these are also
    inflected by gender and case. For example, the verb eiti ("to
    go, to walk") has the active participle forms einąs/einantis
    ("going, walking", present tense), ėjęs (past tense), eisiąs
    (future tense), eidavęs (past frequentative tense), the passive 
    participle forms einamas ("being walked", present tense), eitas 
    (“walked” past tense), eisimas (future tense), the adverbial 
    participles einant ("while [he, different subject] is walking" 
    present tense), ėjus (past tense), eisiant (future tense), 
    eidavus (past frequentative tense), the semi-participle eidamas 
    ("while [he, the same subject] is going, walking") and the
    participle of necessity eitinas ("that which needs to be 
    walked"). The active, passive, and the semi-participles are 
    inflected by gender, and the active, passive, and necessity 
    ones are inflected by case.

The definition you provide here does not apply to Latin.

> every conjugation of an active verb has a corresponding passive form

If we're talking about individual verbs, this does not and cannot apply to any language; there are always verbs that for semantic or syntactic reasons cannot take passive forms. (Examples in Latin: "esse" [to be] has no passive because the passive form of "to be" doesn't exist even conceptually; "sequi" [to follow] has no passive because sequi is, from a strictly syntactic perspective, already inflected as if it were a passive verb.)

If you're talking at a more general level about whether the language offers the possibility of a passive form for every active form a verb might possess, that still isn't true for Latin -- and in fact, the very example you give is quite badly wrong.

> Latin actually had a full 3-dimensional matrix of conjugations: time-based tense (past, present, future), goal-based tense (completed or ongoing), and action (active or passive), with accompanying participles for each case.

Side note: in the traditional analysis, a Latin verb is inflected in five dimensions ("tense", "voice", "mood", "person", "number"). What you call "goal-based tense" is traditionally just called "tense", though I agree with you that it is a separate dimension and shouldn't be conflated with actual tense. In linguistics, this dimension is called "aspect".

Latin participles do not inflect for person or number. Despite what you say here, they also don't inflect for aspect. They do inflect for tense and voice, but of the six possibilities (past / present / future) x (active / passive), only four are possible in Latin -- the present passive and past active forms do not exist for any verb. [1]

(It's possible to make the argument that Latin participles do inflect for aspect, in that some participial forms (the present and the future passive) use the imperfect verb stem and others (the perfect and the future active) use the perfect verb stem. That makes the argument you're making here weaker -- instead of Latin offering four out of six participial forms, it would be offering four out of twelve.)

[1] But remember how "sequi" couldn't have a passive form because it's already formally passive? This opens up a possibility for the perfect participle: secutus can mean "having been followed", the normal perfect passive sense, but it can also mean "having followed", in an otherwise impossible perfect active interpretation of the participle. The "passive form, active meaning" state of the verb can be carried through to its participle.

Italian borrowed all these features from Latin, since it is a directive descendant, tough in Italian these are less strict rules, and nowadays a lot of people doesn't know them and get them wrong (including a lot of politicians, that always get conjunctives wrong).

Moreover, case decensions are very difficult to learn for people from non-case languages. They often exist in addition to verb conjugation for person/number/gender and tense. They have their own divisions into various groups, and irregularities.

Moreover, there are other ways to achieve flexibility in word order, such as particles which agglutinate without tweaking the target word in any way:

Japanese:

PetrがJanaを好きだ。(Petr-ga Jana-o suki da)

JanaをPetrが好きだ。(Jana-o Petr-ga suki da)

Petrが好きだ...Janaを。(Petr-ga suki da, ... Jana o.)

The grammar here isn't standard. Usually you'd mark the subject doing the liking with は and the object being liked with が. Both generally function as topic markers in other circumstances, so context is important. You can even occasionally use は to mark the object being liked if you're pointing out one particularly liked thing. Using を to mark a subject of affection is pretty rare and regarded as incorrect/usually just a tongue slip.

Oftentimes in conversation, particles will be completely dropped in a 好き statement. So getting something like "俺、これ好き" (I, this like) isn't uncommon. In that situation, it's the pause marking the subject doing the liking. If the order is reverse, usually the object marker is used, though as in "これすき、俺は" (this like, I).

Japanese grammar is a complex mess. It uses particles and can be orderless, or it can drop many particles (sometimes all) and have everything marked by pauses and timing.

Heck, you can do that in English sometimes if you pause correctly and don’t mind sounding Shakespearean.

> I love you.

> You, I love.

> You, love I.

> Love you, I do.

> Love I, you.

Your example works because English pronouns are the only nouns in English which still have different forms for parts of speech (e.g., nominative I, possessive my, objective me; he, his, him; they, their, them). In other languages all nouns have different forms for different parts of speech.

In languages like Latin, adjectives also have different forms which match the nouns they modify, making word order flexible without being ambiguous. In English, adjectives usually must preceed the nouns they are attached to, save a few exceptions (attorneys general, tacos supreme, Optimus Prime).

You can do it with surnames as well:

Peter does give his love to Jana.

To Jana does Peter give his love.

Give his love to Jana, does Peter.

It sounds pretty unnatural, but it is understandable I believe. Not sure if the Czech versions would sound natural to a Czech or not.

You've gotten a benefit from having an indirect object expressed with a preposition, but presumably a lot of verbs don't take an indirect object and don't have a synonym that can.

> attorneys general, tacos supreme, Optimus Prime

Apparently "Prime" is a title or rank among the Transformers:

https://en.wikipedia.org/wiki/List_of_Primes_and_Matrix_hold...

In that case, it's probably not best analyzed as an adjective, unless we want to analyze "Senator" as an adjective in "Senator Wyden" or "Mister" as an adjective in "Mister Rogers".

> [...] and don’t mind sounding Shakespearean.

Yes, exactly. When Shakespeare was around, English still had more of these features, and is gradually losing them over time. Shakespearean English always reminds me of German grammar.

The benefit in communication is that the remaining basic form is unambiguous even without proper pausing.

> Shakespearean English always reminds me of German grammar.

One of the strongest phenomena that produces this resemblance is that in Early Modern English, there was no restriction¹ on forming questions by putting the verb before the subject noun phrase (optionally with a question word before both). In modern German, there's still no restriction on forming questions this way.

  Essen Sie Fleisch?
  eat   you meat?

  Sprechen Sie Französich?
  speak    you French?

  Kommt Ihr aus  Russland?
  come  you from Russia?

  Glaubt   er an Gott?
  believes he in God?

  Wohin   geht sie?
  whither goes she?

  Wie macht er das?
  how does  he that?

  Wir sind hier - wann kommt Ihr?  [a travel ad slogan]
  we  are  here - when come  you?

In modern English, there is a new restriction that the verb that gets moved to the front can only be a form of "do"², "be", or an auxiliary verb (e.g. "can", "may", "should", "must", "might", "will", or "have"³, among others). When the main verb is not one of these, the sentence must be turned into a paraphrase (almost always using "do") before forming the question, like "I do speak Portuguese" → "Do you speak Portuguese?" or "I did have a pet lizard" → "Did you have a pet lizard?".

I have a German colleague who doesn't seem to have learned this restriction and continues producing German-style, or Shakespearean-style, questions in writing like "What contains that file?", "What message produced that command?" or "Updated you the configuration recently?".

In one of my college linguistics classes, I tried to deny that this change was real because I could still easily understand the Shakespearean questions. The trouble is that they nonetheless sounded archaic to me and I wouldn't actually have produced them myself in speech. So really, I should have admitted that there was a practical difference between comprehensibility and acceptability (and that difference with regard to a historic form is exactly what "sounding archaic" refers to).

¹ I'm not totally sure that there was no restriction at all; it may have begun to appear already but just not been complete.

² Though not as a main verb! Shakespeare has "wherefore did you so?" in Macbeth and "wherefore didst thou so?" in King John, which in modern English would have to be "why did you do so?" or "why did you do that?" rather than "why did you so?" or "why did you that?".

³ In some varieties of English "have" can be used this way in both the possessive and auxiliary senses ("have you any wool?" / "have you eaten?"), but in other varieties only in the auxiliary sense (in which case the paraphrase with "do you have..." is required when asking about possession). I'm not sure whether this still works in the past tense, or only in the present tense.

Excellent write-up!

Btw, in modern spoken German we mostly only have two tenses left: present and perfect. Present is used for the present and the future. Perfect (formed with 'haben') is used for past events.

(Yes, we also still use other tenses. But they are much rarer in spoken German.)

And especially in spoken questions about things people did or didn't do, an auxiliary verb (like "haben" or "sein") is almost mandatory.

"Gingst Du nach Hause?" vs "Bist Du nach Hause gegangen?"

Ok, perfect excuse for me to wheel out the Buffalo! https://en.wikipedia.org/wiki/Buffalo_buffalo_Buffalo_buffal....

Also note that poetic forms and loosened syntax go and in hand.

Can you explain how your first two examples mean the same thing?

Shameless plug, I wrote about this here: https://blog.mclemon.io/cases-are-hard-who-punched-who - I used a slightly more one-sided example (Petr punches Pavel) to start things off ... but I talked a bit more about other cases to explain how it all fits together and why they're important.

Basically though, in the specific example in the GP comment we know who loves whom by the order of words in the sentence - "${subject} loves ${object}" which is different from "${object} loves ${subject}". We know this because the word order in English tells us this.

But in Czech and Latin this isn't how it works - the word order doesn't have the same significance or meaning as it does in English [1]. You indicate who does what by making a little modification to the words. In this example it's one of the simpler cases - a simple verb with a subject and an object. The subject is the "nominative" case (like the dictionary/default form of a word) and the object should be in the "accusative" case. So it's actually something like "${object.nominative()} loves ${subject.accusative()}".

And just to confuse you further - it's not just "who does what" - there are other cases that denote doing something with or to or in :-) I give some examples in the above link

[1] = that's not to say word order doesn't have a role, it's just not the same role as in English

Regarding the linked blog page: It's a nice introduction to the idea of cases, but it's not entirely accurate: the usual dative/locative form of Pavel is Pavlovi. Pavlu would be used only when there's a surname present. Thus "o Pavlovi" but "o Pavlu Novákovi".

Oops, thanks! Fixed - these nuances still kinda escape me :-O

> ${object.nominative()} loves ${subject.accusative()}

You mixed it up, it should be "${object.accusative()} loves ${subject.nominative()}".

Or, I guess, how you define "object" and "subject" for "loves" :P But grammatically, for transitive verbs, "subject" is the doer of an action, "object" is the target of an action.

Damn I rewrote that comment about 2 times and must have mangled it in the process!

I assume "Janu" is the accusative form of "Jana", indicating that Jana is the recipient rather than the source of the love in each sentence.

That assumption is correct. You see the same conjugation form in other Slavic languages, including Russian.

Nouns/names/numbers have declensions. Verbs have conjugations.

I think the coverall term that btilly wanted was "inflections".

Extremely interesting and extremely frustrating. I'm entering my fourth month learning Czech. Thanks for the comment.

    Janu miluje Petr. -- PETR loves Jana.
    Petr miluje Janu. -- Petr loves JANA.

Does context define subject/object since otherwise these seem equally to mean Jana loves Petr?

These sentences do not mean "Jana loves Petr." Subject/object relationship is determined by case marking suffixes on nouns.

Another comment proposed an apropos analogy:

English is like position-based parameter-passing.

  sentence.love("Petr", "Jana")

Czech is like keyword-based parameter passing.

  sentence.love(object="Jana", subject="Petr")