Parts of Speech (semantics vs. syntax; pronouns, numerals & adjectives, nouns) + discussion

By Godmy, in 'General Latin Chat (English)', Jun 9, 2018.

  1. Godmy A Monkey

    • Civis Illustris
    The Argument:

    The definition of Parts of Speech in Latin/in general

    A Short Introduction:

    In linguistics we use different terms for different kinds of a description of a language. The different descriptions are different levels of abstraction = different perspectives of describing a language. For each of these different levels we use different terms and the terms from one category/level should not describe elements from the other level (or category) and vice-versa in order for the definitions to have some inner consistency.

    The most salient levels of abstraction in linguistics are (from the simplest to the most complex):
    • phonology
    • morphology
    • syntax
    • semantics
    (There are many more levels than that, some are mid-steps in between the mentioned ones (e.g. morphosyntax, nano-syntax), some are not as easily put into this order (like lexicology)... and many others I haven't mentioned (pragmatics... etc.) )

    Here I will focus solely on syntax and semantics (and I touch superficially on morphology too) as to talk about Parts of Speech.


    The terms used in Syntax (+ a short definition):
    • subject
    • object
    • attribute
    (And other terms I won't be naming (adverbial, predicate, complement, clause, phrase, sentence element, direct, indirect, syntactic role ...etc.), these are the most salient ones)

    subject - a phrase in the clause (a unit consisting of one or more words) that connects on the equal level only with the predicate which makes a statement about this unit. The nominative case[morphology] usually correlates with the subject[syntax].

    object -a phrase/a unit in the clause which connects itself to the head of the predicate (usually the verb) and complements the meaning of the predicate. The accusative[morphology] usually correlates with the object[syntax] if the object is direct.

    attribute - a phrase (more often one-worded) which is a part of a noun phrase[tree/phrasal syntax] which itself is either a subject, object or complement. The genitive case[morphology] and several parts of speech often correlate with the attribute, very rarely in Latin a prepositional phrase[tree/phrasal syntax] too.


    The terms used in Morphology:
    • declension
    • conjugation
    • mood
    • morpheme
    • morph
    • case
    • case-ending
    • gender
    • number
    • finite
    • non-finite
    • (etc.)

    The terms used in Semantics:
    • agent
    • patient
    • semantic roles
    • *nouns, adjectives, pronouns, verbs, adverbs, prepositions, particles, interjections, (participles**), (determiners***)
    • (etc.)
    *I will say why
    **According to some Roman grammarians, but the modern grammarians use term rather morphologically - as a kind of verbal mood (i.e. finite moods: indicative, imperative, subjunctive, [optative]; non-finite moods: infinitive, participle)
    ***Used sometimes as a part of speech[semantics] by modern English grammarians for their own special group of words in their NPs (noun phrases), but it has some implications for syntax too so it's a hybrid (semanticosyntactic) term and perhaps unfit/inept for Latin while apt/fit for English, I will explain later why.


    Sorting the PARTS OF SPEECH as linguistic terms to one of the mentioned linguistic categories:

    The best definition for the terms from the "Parts of Speech" group (i.e. in Latin "noun, adjective, pronoun, verb, adverb, preposition, particle, interjection,") is such where the terms don't copy the already existing terms (their definitions) from some of the other categories. For example, if we somehow said that a noun is a subject, then we don't need the term subject anymore or vice-versa, we don't need the term "noun" anymore; if we said that a noun is a subject or object, we don't need the term "sentence element" anymore, the same applies for attribute (e.g. noun in the genitive case), complement and so on and so on.

    The same as when we were able to introduce the semantic roles such as agent and patient and make them useful by being distinct/independent on their occasional(!) syntactic counterparts such as subject and object, the definition of any individual part of speech should be independent on which syntactic role the part of speech takes in a clause, it should be distinct.


    A noun can be [syntax] an attribute(= a "describer" if in the genitive case; That is the head of an attribute phrase, a tail to the noun phrase), a noun can be a subject (+respective phrases), it can be a subject, object, complement. But a noun will consistently be as defined by semantics "a word that refers to a person, place, thing, event, substance, quality, quantity...".

    An adjective can be [syntax] an attribute (a "describer") (+the respective phrases), sometimes it's the complement(predicate), it can sometimes even stand for the whole subject and object (subject & object phrases), yet even in such cases no confusion arises, since we distinguish syntactic roles from the semantic description. We don't mind that an adjective (unless it lexically started to be as a full noun ) stands for a subject, since it's a mere semantic definition and the syntax of the clause stands unshaken. The definition of an adjective will consistently always be (no matter its current syntactic role): "a word that refers to a property or relation of nouns"

    A pronoun can take various syntactic roles[syntax]. It can be a subject (+respective phrases...), object, complement or attribute ("a describer"). No matter which syntactic role it takes in a sentence, whether it connects directly to the predicate[syntax] as subject, whether it directly complements a verb[semantics] as object[syntax], whether it describes[syntax] either subject or object or complement, the term is still useful, since it has always one consistent and unshaken definition which could be simplified as: "a word where the meaning of the pronoun is dependent on an antecedent [or an implicit lack thereof]"

    A numeral according to some of its multiple kinds can similarly to pronoun take various syntactic roles (subject, object, complement, attribute), but no matter which syntactic role it takes in the sentence, the term "numeral" is still useful since it has its consistent and never changing definition: "a word characterized by the designation of numbers"

    Conclusion: Based on this the argument of the usefulness of terms, the only case where they have a consistent definition which hasn't been used by any other linguistic category for any other term (& therefore are useful as standalone linguistic terms, be distinct) is when the parts of speech are put into semantics. The semantic definition complements nicely with the syntactic role of the same word and the definitions don't overlap or "fight themselves" or "aren't conflated" in any way. If the parts of speech are put to any other category but semantics (albeit temporarily), we introduce strange terms into e.g. syntax, such as "person, place, property, antecedent" and so forth which have no place in syntax, or in such case we just involuntarily copy the definitions such as those of subject or object just to rename them "nouns" or "adjectives", which is inelegant from the said reasons. (I will speak further about the determiner).


    Categorization of Pronouns:
    • Personal pronouns (prōnōmina persōnālia)
    • Possessive pronouns (prōnōmina possesīva)
    • Demonstrative pronouns (prōnōmina dēmōnstrātīva)
    • Interrogative pronouns (prōnōmina interrogātīva)
    • Relative pronouns (prōnōmina relātīva)
    • Negative pronouns (prōnōmina negātīva)
    • Indefinite pronouns (prōnōmina indefīnītīva)
    • (+ sometimes other special categories)*
    *e.g. redditiva, redistributitiva.. etc.

    Personal pronouns (prōnōmina persōnālia): (such as here) ego, tū, nōs, vōs (*ea, is, id/ille, illa, illud) + the respective cases of each.

    Syntactically the personal pronouns usually take the role of a subject ([Ego] currō.), object (Mē vidēs.), complement (Ego sum, quem vidēs.) or attribute in the genitive case[morphology] (Parva pars meī vītābit Libitīnam).

    *They often semantically stand for the personal pronouns, in case they are needed at all, but they are semantically rather demonstrative in Latin.


    Possessive pronouns (prōnōmina possessīva): (such as here or also here...) meus, tuus, noster, vester, suus + respective cases.

    Syntactically the possessive pronouns take the role of an attribute ("a describer") (Canis meus mortuus est.), less frequently even the syntactic role of a subject (Nostrī vīcērunt.) and object (Nostrōs occidērunt).


    Demonstrative pronouns (prōnōmina dēmōnstrātīva): (such as here or here[scroll up a bit too]) hic, haec, hoc | is, ea, id | ille, illa, illud | iste, ista, istud | tālis, tāle | īdem, eadem, idem | ipse, ipsa, ipsum | sōlus, sola, sōlum | *ūnus, ūna, ūnum + their cases

    Syntactically the demonstrative pronouns can take almost any role: subject (Ille, quem digitō ostendō, vēnit.), object (Illum, nōn hunc, videō.), attribute (Ille canis mortuus est.), complement (Ille est, quī vēnit.).

    *This is a demonstrative pronoun only in the sense "sōlus".

    Interrogative pronouns (prōnōmina interrogātīva): (such as here or here) quis, quid, | *quī, quae, quod | *quālis, quāle | quantus, quanta, quantum** | uter, utra, utrum...etc. + their cases

    Syntactically the interrogative pronouns can take the role either of subject (Quis vēnit? Quid factum est?), object (Quid vīs? Quem vīs - quid nōmen eī est? ) or attribute (Quem discipulum maximē amās, magister?)

    *These are interrogative pronouns only when in a direct or indirect question. For relative clauses, see relative pronouns.
    **If not the "quantum" adverb.


    Relative pronouns (prōnōmina relātīva): (such as here or here) *qui, quae, quod | quīcumque ... etc. + respective cases

    Syntactically the relative pronouns can mainly take either the role of the subject (Equus, quī currit, fessus fit.) or the object (Equus, quem videō, pulcher est.), attribute (Equus, cujus color albus est mihi placet), indirect object (Equus, cui dentēs nōn sunt, mihi nōn placet.)

    *These are relative pronouns, unless it's the introducing word of a direct or indirect question.

    Negative pronouns (prōnōmina negātīva): (such as here or here) nēmō | nūllus, nūlla, nūllum | neuter, neutra, neutrum ... + their respective cases.

    Syntactically the negative pronouns can take the role of a subject (Nēmō vēnit.), object (Nūllum videō.), complement ([Ille] nēmō est), attribute (Nūllum canem videō.)

    Indefinite pronouns (prōnōmina indefīnītīva): (such as here or here) aliquis, aliquid | aliquī, aliquae, aliquod | ūllus, ūlla, ūllum | tōtus, tōta, tōtum | quīdam, quaedam, quoddam | quiddam | alius, alia, aliud | *ūnus, ūna ūnum... + their respective cases

    Syntactically the indefinite pronouns take various syntactic roles like subject (Aliquis vēnit.), object (Aliquem videō.), attribute (hunc sī ūllus deus amāret... / ibi ūna mulier aderit lepida)

    *This one is indefinite only in the same "someone"

    I'm not adding a treatise on numerals at this moment, since this one is long enough.



    A determiner is usually a term used in a syntactic relation: a phrase which is a necessary part of a noun phrase. This is how English treats this. Whenever the English NP happens to have no determiner, we say the determiner is zero. Therefore we treat determiner as an ever-present part of the noun phrase. (in the same time the term, unfortunately, happens to be used also in parts of speech in English, therefore in semantics, creating a hybrid). Should we import these syntactic features to Latin, we must say that Latin just like Czech and most Slavic languages (with two exceptions) has no determiner, since the Latin NP has no determiner in the same way.

    I would also try to stay away from a general/international/non-controversial look at the Latin grammar that must introduce a new term that has never been part of the grammatical tradition begun with Latin in the West (albeit having its root in Greek), but that's just my subjective argument.



    I'm not sure Pacifica, this post is well-researched and sourced enough for your standards at Grammar Tips, but in any rate, I spent a few hours on making it, formatting it, I also included links to various sources (none of them Czech, even though that would have been easy). One doesn't need to agree with that argument as I laid it down, but it is valid/innerly consistent in respect to linguistics (+ supported by the grammatical tradition quite alive still in number of culturally western countries) and mainly, it's not "my" fiction, I paraphrase in the argument what is taken as a pretty wide consensus and has been for centuries amongst plenty of grammarians. But whatever.

    Ser I added my bit on the determiner.

    AoM I'm willing to use for simplification (and I do believe that OLD simplifies and in a way A&G too with its loose use of terms such as "syntax" and in a way even "adjective") the words as "adjectival" or "substantival" to mean "syntactically behaving like subject or object" for substantival and "syntactically behaving like some attribute" for adjectival, that's why I was willing -with the highest reluctance- to grant terms as "adjectival pronoun" e.g. And I would grant terms such as "syntactic noun" or "syntactic adjective" for different uses of e.g. ille, but that already induces, in my view, too much of a terminological confusion and a category conflation, so, even though I might have used the terms in the past in different threads, considering it a lesser evil, I think I will try to abstain from that too. I stand behind what I wrote in the "argument" part of this post, since that makes sense to me.

    Other people interested in this discussion (since I don't want to hijack the grammar thread anymore, since I do think it should be for the beginners who might be now dissuaded to post there anymore - I hope not) were Hemo Rusticus, Dantius, indirectly also Lysandra who might be interested in the post (in "not just mine" take on parts of speech, especially pronouns).
    Last edited by Godmy, Jun 11, 2018
  2. Hemo Rusticus Tom Bombadillo

    • Civis Illustris
    Gadzooks, man. All this pedantry and intellectual masturbation (that is to say, re-writing what any rube can find in an elementary grammar) because you fail to understand that new-comers to Latin have no interest in semantics to rival your own?

    Let me go see what Dunning and Kruger have to say about this kind of thing.
  3. Pacifica grammaticissima

    • Civis Illustris
    Tone down, Hemo... Godmy has obviously spent quite a lot of time and effort to explain to us what he meant.

    I understand your point, Godmy, and thanks for your post. However, I still think that it isn't necessarily the only valid way of looking at things. Maybe I will think some more about it and change my mind someday, who knows. But for now this is my stance.
    Godmy likes this.
  4. Godmy A Monkey

    • Civis Illustris
    Hemo Rusticus I don't understand what you wanted to say to the topic? If you have a counterargument say it, otherwise it's time wasting.

    I think this was worth it, because it's obviously unresolved topic amongst many Latinists and I want to use the post in the future as a reference.
    Last edited by Godmy, Jun 10, 2018
  5. Godmy A Monkey

    • Civis Illustris
    Also I quite don't understand "intellectual masturbation" (are you projecting?). If you've ever written a paper on university, this is quite standard... Or is too long for you?
    Last edited by Godmy, Jun 10, 2018
  6. Hemo Rusticus Tom Bombadillo

    • Civis Illustris
    University? What are that?

    Look, pal-o'-mine, nobody likes being called 'wrong', least of all semanto-wizards and historical linguists like you and me. I only question the value of this re-hash of material that is (more than likely) available elsewhere on this forum. Why do you think Puella Pacis redirects people to posts that she herself has already tailored for the issue at hand?

    In short, you've conceded that a term like 'adjectival pronoun' for 'meus' would "make at least some sense" for you, whose definition of these categories is founded on semantics, which as we've been reminded many a time 'is always the king'.

    I, and many others, whose definition is founded on (historical) morphology, prefer the term 'pronominal adjective'. Isn't there some way to split the difference? Some twenty-part hyphenated monstrosity that will make every semanticist, semioticist, statistician, metaphysician, and new-wave musician happy? Adjectivo-pronomino-phenomenonico-particulo-meta-phonemic?
  7. Godmy A Monkey

    • Civis Illustris
    It's not available elsewhere on the forum. Otherwise I wouldn't have bothered. There is not a single a word on the forum I'm aware of which emphasizes this or contrasts with what is written in some of the mentioned English resources with how it *should* (my opinion) rather be defined (although I remember I complained about this at least once briefly years ago), I've been here for 8 years, so I have some idea.

    I structured this post this way so it could potentially serve as a similar resource as those which are posted to the "Grammar Tips" moderated section (= so it's as readable as if from a textbook). I wanted it to be rather long so everything I say - every term is explained and it's clear even for people/Latinists not much interested in any linguistics at all...


    What is then "historical morphology" in here? Like, anything which declines and has some xy features is an adjective? Or how do you determine it?
    I just don't see how you and others would define it. And that's why here's this thread, so we can lay the argument in peace and not interfere with beginners exercising their Latin in there...
    Last edited by Godmy, Jun 11, 2018
  8. Iáson Cívis Illústris

    • Civis Illustris
    Not that I have particularly strong views either way on the matter, but this is what Servius says:

    dē prōnōmine

    Prōnōmen dictum est, quia pōnitur prō nōmine, ut sīquis dīcat 'Vergilius scrīpsit būcolica, ipse scrīpsit geōrgica.' nec tamen videntur prōnōmina, quoniam nōminum funguntur officiō, esse nōmina. nam etsī nōminum vim exprimant, nōn tamen plēnē exprimunt. nōmina enim posita plēnās faciunt ēlocūtiōnēs, prōnōmina sēmiplēnās. item nōmina posita velut inchoant ēlocūtiōnēs, prōnōmina autem velut conplent. quālitās prōnōminum prīncipāliter bipertita est. omnia enim prōnōmina aut fīnīta sunt aut īnfīnīta. fīnīta sunt quae recipiunt persōnās, id est quae dēfīniunt persōnās, et sunt tantum tria, ego tū ille. sed ille plērumque variat. nam sī ad praesentis persōnam referātur, tunc vērē fīnītum est; sī dē absente dīcātur, minus quam fīnītum est. īnfīnīta autem prōnōmina variās habent speciēs. nam licet omnia quaecumque nōn recipiunt persōnās īnfīnīta sint, tamen alia dīcuntur generāliter īnfīnīta, alia minus quam fīnītā, alia articulāria vel dēmōnstrātīva, alia possessīva. generāliter īnfīnīta sunt quae cuicumque persōnae aptārī possunt, ut est quis; et sunt haec septem tantum, quis quālis tālis quantus tantus quotus tōtus. minus quam fīnīta dīcuntur quae commemorātiōnem habent notārum personārum, ut est ipse: haec sunt sex tantum, ipse iste is hicc īdem suī. sed ex hīs alia sunt quae absentēs persōnās significant, alia quae praesentēs: absentēs reliqua omnia; praesentēs haec tantum, iste ista istud et hic haec hoc, quae nōn nūllī prōnōmina etiam articulāria vocant, eō quod mōre Graecō cum nōminibus dēclīnantur. possessīva dīcuntur prōnōmina quae nōs aliquid possidēre ostendunt, ut est meus tuus. haec in quattuor partēs dīviduntur. aut enim utrāque parte singulāria sunt, aut utraque plūrālia, aut intrīnsecus singulāria, extrīnsecus plūrālia, aut intrīnsecus plūrālia, extrīnsecus singulāria. utrāque parte singulāria sunt, ubi et persona possidentis et possessī ūna est, ut meus; utrāque parte plūrālia, ubi multī sunt quī possidentur et multī quī possident, ut est nostrī; intrīnsecus singulāria, extrīnsecus plūrālia sunt quae possidentis ūnam persōnam ostendunt, possessōrum plūrimās, ut est meī, vel ē contrāriō, ut est noster. intrīnsecus autem dīcitur quod pertinet ad persōnam possidentis, extrīnsecus quod pertinet ad persōnam possessī. sunt autem possessīva prōnōmina quīnque tantum, meus tuus suus noster vester. absque hīs prōnōminibus, quae omnia sunt vīgintī et ūnum, nūllum aliud, quod suam orīginem habeat, poterit reperīrī; sed sīquā sunt, conpositiōne fīunt.
    Scīre autem dēbēmus in dēclīnātiōne prōnōminum plērumque sub ūnā significātiōne cāsūs variē prōferrī, ut est quis vel quī: nam possumus dīcere 'quis tibi fēcit iniūriam' et 'quī tibi fēcit iniūriam'. item invenītur genetīvus duplex, ut meī vel mīs, tuī vel tīs: nam possumus dīcere 'meī causā tē petō' <et 'mīs causa tē petō' et 'tuī causā tē petō' et 'tīs causā tē petō'>. sed mīs et tīs dē ūsū recessērunt. item ablātīvus singulāris aliquandō variē terminātur, ut ā quō vel ā quī. nam dīcimus 'ā quō vēnistī' et 'ā quī vēnistī'; sed ā quī in ūsū esse dēsiit. dativus etiam et ablātīvus plūrālis plērumque variī sunt: dīcimus enim et ā quīs et ā quibus. quī quidem videntur hāc ratiōne variārī, quoniam ablātīvus singulāris inventus est varius. scīmus enim quia omnia quae ō terminantur in īs mittunt, quae ī, in bus. ergō sīcut dīcimus ab hōc doctō ab hīs doctīs, ita dīcimus ā quō ā quīs. item sīcut dīcimus ā puppī ā puppibus, ita dīcimus ā quī ā quibus. sed prōnōmina quem ad modum in dēclīnātiōne duplicant cāsūs, ita plērumque minuunt. nam sunt prōnōmina ubi tantum modo genetīvī sunt * * * ut eccum eccam ellum; sunt ubi nōminātīvus et vocātīvus deest, ut suī sibi sē ā sē; sunt ubi vocātīvus tantum deest, ut in prōnōmine quod est ego.
    Godmy likes this.
  9. Hemo Rusticus Tom Bombadillo

    • Civis Illustris
    I guess I found that shocking, but as a veteran of the forum you know more about its guts than I do, as you said.

    And your opinion is very welcome here, but when you mention it matter-of-factly, and wag your finger at us saying 'you and the tradition you've learned is wrong', your audience will, understandably, resist.

    Yes, more or less. Only, from a historical perspective, words like solus lack a key morphological feature that 'true pronouns' can have. Take for example the multiple reflexes of the Indo-European interrogative pronoun with different vowel-grades ('ablauts'):

    kwos – kweH2 – kwod > L *quī quae quod

    kwis – kwid > L quis quid

    (< Proto-Italic */kwoi kwai kwod/)

    Edit: The 'w' in each 'kw-' is supposed to be superscript, but the distinction isn't too important right now.
  10. Godmy A Monkey

    • Civis Illustris
    Thanks, Iáson, I will read it later in the day! It looks very interesting!

    Hemo Rusticus you misunderstood: my opinion is that it should be so defined also here, on this forum, when communicating to beginners (and I think that's reasonable). But it's not my opinion that it is, as a matter of fact, defined like that in most places I know, most textbooks I came in contact with in different kinds of schools, with all grammar teachers I've ever had. It's not just "my tradition". If you look up the links I provided in the post, you can see Italian, German, French and plenty of Latin resources for this (I purposely avoided Czech). I sometimes feel as if you wanted to picture me as some outlier with ridiculous views... I agree that it is my view that the forum community takes a strange stance on this, but all that more I feel it's important that somebody points it out (that's also my opinion, lol).

    How do you then distinguish a noun from an adjective, since they both decline? I really want to hear from some of you a clear definition of a noun, adjective and pronoun which in the same time allows you to put there meus (into adjectives) and in certain cases e.g. "ille". I already took the pains to provide my definition. (when I say "my", I don't mean that I'm the author, obviously)

    Hm, but you say in one sentence both "true pronouns" and "can have". So those that don't have it are true pronouns too.. or not? And how does the morphology perfectly determines a pronoun? I'm just confused here.
    Last edited by Godmy, Jun 11, 2018
  11. Hemo Rusticus Tom Bombadillo

    • Civis Illustris
    I believe you, neither is it the case that what I'm explaining is merely "my (own) tradition".

    I say 'true pronouns' for the sake of the argument; these are words that you and I would both agree (I think) are unarguably pronouns, like qui.
    & I say 'can have' because we only have attestations of some of these reflexes (which doesn't mean the unattested ones never existed).

    You've mentioned that an adjective cannot have an antecedent, that it is rather pronouns that do have them, and that solus is in fact a pronoun. Can you give an example of solus used with an antecedent, so I may understand better?

Share This Page


Our Latin forum is a community for discussion of all topics relating to Latin language, ancient and medieval world.

Latin Boards on this Forum:

English to Latin, Latin to English translation, general Latin language, Latin grammar, Latine loquere, ancient and medieval world links.