Comparing Greek Morphologies – Part II

Parts of Speech – Determiners, Particles, Pronouns & Others:

In what discussion that follows, we will examine our five morphological databases’ parts of speech beginning with the Friberg, which as the fewest categories and ending with Swanson, which as the most. In this post we will focus on the Article, Particles & related categories, and pronouns.



The Friberg database consists of the fewest number of categories with eight. What is unique about Friberg, other than the number of categories, is the use of the term “Determiner” rather than article, though it seems that beyond the difference in name, the analysis is the same. Only the Greek article is included in “Determiner.” Other word classes that could fall into that category, such as demonstratives, are analyzed as adjectives. And the category “Adjective” has some surprising words, particularly relative pronouns. Their explanation for this is,

Relatives function as part of the adjective system in our analysis for two reasons. First, whole relative clauses usually function to modify a noun in the same way an adjective does. Second, a few relatives are simple modifiers (a-r instead of apr) of following nouns. Among them are the ἣν found in Matthew 10.11 (a-raf-s), and the οἵους found in 2 Timothy 3.11 (a-ram-p). Because relatives work analogously to adjectives, they are appropriately placed in the same category.

I take issue with this for a couple reasons; one, historical and the other, syntactic. Historically, BDF is helpful here:

The original use of ὁ ἡ τό as a demonstrative pronoun is retained in classical usage in certain fixed phrases; the forms of the old relative pronoun ὅς ἥ ὅ replace it occasionally in classical and more frequently in Hellenistic times. The origin of this confusion was, on the one hand, the old sigmatic alternative form of ὁ: ὅς which in Greek had become identical with the relative in form; and, on the other, the Epic and dialectal use of ὁ ἡ τό as a relative pronoun (cf. the article der in German which serves as article, relative and demonstrative; in English that is both demonstrative and relative and is related to the article). Cf. K.–G. ii 227. In the NT (except the Epic quotation from Aratus in A 17:28 where τοῦ = τούτου) there are preserved only ὁ μὲν … ὁ δέ (ὃς μὲν … ὃς δέ) ‘the one … the other’ and ὁ δέ ‘but he’, ὁ μὲν οὖν ‘now he’. Other expressions like καὶ ὅς (Homil Clem 6.2.13 καὶ ὃς ἔφη), καὶ τόν ‘and he, him’, τὸν καὶ τόν ‘such and such’, or ‘so and so’, πρὸ τοῦ ‘formerly’ have completely disappeared. (BDF §249-251)

Furthermore, relative pronouns are syntactically distinct from adjectives. They do not have the same distribution. Adjectives, when modifying an articular noun, can appear on either side of that noun. Relative pronouns can only follow articular nouns. Searching, a relative pronoun precedes its head noun only when that noun does not have the article and its not common.

Demonstratives should not be considered adjectives either. They are syntactically distinct. The distribution of the demonstrative compared to the Adjective in continuous noun phrases shows this easily:

Adjectives appear in the following positions:

Art (Adj) N (Adj)

But Demonstratives appear elsewhere:

(Dem) Art N (Dem)

Together the differences are accentuated:

(Dem) Art (Adj) N (Adj) (Dem)

Demonstratives will always appear further away from the head noun than adjectives unless that adjective is also articular and even this is quite rare (cf. Matt 12:45; Mark 12:43; Luke 21:3)

For Friberg, Particles include negators, certain modals, and a variety of other words. Doing a search through the NT and organizing it by lemma, we find that all conditionals are label as particles, as is ἰδού, ἀμήν, ἂν, γέ, and so on. I am somewhat perplexed by the fact that 18 times γάρ is label as a particle and twice labeled as both a particle and a conjunction.

Also of note for Friberg is the lack of a category for Indeclinable/Foreign words, which is included in every other database examined.



Logos uses eleven “parts of speech.” Helpfully, they include “Indeclinable word.” Demonstratives and Relatives are included in the category of “Pronoun.” I consider this a step in the right direction from Friberg, but it stops short. Functionally and semantically, Demonstratives are more like pronouns than they are Adjectives as in Friberg in that they are referential, but at the same time, they once again are syntactically distinct from pronouns.  This is especially true of demonstratives, which syntactically appear on the extremity of the noun phrase. If we look at the distribution of demonstratives and pronouns separately in standard continuous NPs (i.e. no hyperbaton), it is easy to they the differences. Pronouns appear here:

Art. (Pro) N (Pro)

But demonstratives appear here:

(Dem) Art. N (Dem)

When we put them together in the same NP their distributions look like this:

(Dem) Art. (Pro) N (Pro) (Dem).

Syntactically speaking, we cannot consider them to be in the same class of words.

I find the category “interjection” interesting, particularly since it only seems to be used 70 times in the GNT by Logos. ἔα, οὐά, οὐαί, ὄφελον, and . Once, ἕν receives the label, which might be an error. I question the helpfulness of the category when its use so few times. Even it Josephus its used only 241 times and Josephus is a quite large corpus. Other than that, Logos’ category “Particle” is quite similar to Friberg’s, though it does not include the conditionals εἰ and ἐάν (it does include ἂν). I think this is reasonable since syntactically, the former introduce subordinate clauses, whereas ἂν does not. I would likely do the same. It also seems to include certain connectives such as μέν as “correlative particles” rather than conjunctions. Both Friberg and Logos have the option of “Interrogative Particles.” For Logos it seems to include mostly negators and strangely a few copulas.

Finally, for the life of me, I do not understand why Logos (or Robinson or Swanson for that matter) use the term “Definite Article” when Greek only has one article to begin with. And even then it is not even necessarily identical to our English Definite article.



Probably the most striking category in Robinson is the differentiation between “Conjunction” and “Conj. or Cond. Particle.” The latter category consists of the following words: ἐάν (346), ἐάνπερ (3), εἰ (436), εἴ (67), εἴγε (5), εἴπερ (6), κἄν (13). I cannot help but wonder why ἂν is not included here. ἂν rather appears in “Particles.”

When we look at interjections, Robinson’s use of the category is slightly different than that of Logos. They both include a nearly identical set of words:  ἔα, οὐά, οὐαί, and . Logos seems to have added ὄφελον to the lot, but that’s it.

The words labeled as interrogative particles are: ἆρα(19), μήτι (16), οὐχί (46), πότε (19), ποῦ (44). Again, the inclusion of the negator is interesting, though I understand this based on BDAG’s third definition.

Robinson also continues in parallel with Logos in terms of its organization of pronoun types, including demonstratives, and also in its label of the article as “Definite.”



Gramcord uses 13 parts of speech. They rightly call the Greek article, “Article.” Their category “Foreign Word” generally parallels Robinson and Logos’ “Indeclinable word.” Interestingly they add “Improper Preposition,” for those adverbs that function as prepositions in the Hellenistic period – I question whether the native speaker would have made such a distinction. Probably most helpfully, they also include a category for punctuation – something not seen elsewhere in the databases I have available, though I know that it is searchable elsewhere.

Particles include the following labels:


Worth examining, in my opinion, were the categories “Alternating,” “Comparative,” and “Indefinite.” In these categories, there are 495 occurrences, which interestingly enough include the following words:

δέ (66), ἤ (45), ἤπερ (1), καθάπερ (2), μέν (37), ποτε (22), ποτέ (7), που (3), πού (1), πως (15), ὡς (267), ὡσεί (21), ὥσπερ (7), ὡσπερεί (1)

The occurrences of δέ generally seem to parallel μέν. They also seems to always receive two tags: one for conjunction and the other for “alternating particle.” The rest seem to be self-explanatory, though I would be interested in thoughts from others on any of these.

GRAMCORD also provides a high level of specification for pronouns:




Swanson’s Parts of Speech consists of 14 categories. The categories of Substantive, Contraction, and Number/Word are unique to their database. Substantives consist essentially of adjectives being used as nouns.

Words includes in “Contraction”:

κἀγώ (76), καθό (4), κἀκεῖ (10), κἀκεῖθεν (10), κἀκεῖνα (4), κἀκεῖνοι (7), κἀκεῖνον (3), κἀκεῖνος (6), κἀκεῖνός (1), κἀκείνους (1), κἀμέ (3), κἀμοί (5), κἄν (17), τοὐναντίον (3), τοὔνομα (1)

Words included in “Number/Word”:

δέκα (25), δεκαοκτώ (2), δεκαπέντε (3), δεκάτας (2), δεκατέσσαρες (3), δεκατεσσάρων (2), δεκάτη (1), δεκάτην (2), δέκατος (1), δεύτερος (1), διακόσιαι (1), διακοσίας (1), διακοσίους (2), διακοσίων (3), δισχίλιοι (1), δύο (126), δυσί (1), δυσίν (5), δώδεκα (75), δωδέκατος (1), ἑβδομήκοντα (5), ἕβδομος (1), εἴκοσι (11), εἷς (98), ἑκατόν (17), ἑκατονταπλασίονα (3), ἕκτη (3), ἕκτην (3), ἕκτης (2), ἕκτος (4), ἕκτῳ (2), ἕν (71), ἕνα (42), ἐνάτῃ (1), ἐνάτην (5), ἐνάτης (3), ἔνατος (1), ἕνδεκα (6), ἑνδεκάτην (2), ἑνδέκατος (1), ἐνενήκοντα (4), ἑνί (22), ἐννέα (5), ἑνός (33), ἕξ (13), ἑξακόσιοι (1), ἑξήκοντα (9), ἑπτά (87), ἑπτακισχιλίους (1), ἔτη (1), μιᾷ (18), μία (17), μίαν (36), μιᾶς (8), μυρίων (1), ὀγδοήκοντα (2), ὄγδοος (1), ὀκτώ (8), πέμπτος (1), πεντακισχίλιοι (4), πεντακισχιλίους (1), πεντακισχιλίων (1), πεντακόσια (1), πεντακοσίοις (1), πέντε (38), πεντεκαιδεκάτῳ (1), πεντήκοντα (7), πρῶτος (1), τέσσαρα (6), τέσσαρας (7), τέσσαρες (11), τεσσαρεσκαιδεκάτη (1), Τεσσαρεσκαιδεκάτην (1), τέσσαρσιν (5), τεσσάρων (12), τεσσεράκοντα (22), τεσσερακονταετῆ (1), τεσσερακονταετής (1), τεταρταῖος (1), τετάρτῃ (1), τετάρτην (2), τετάρτης (1), τέταρτον (2), τέταρτος (3), τετάρτου (1), τετρακισχίλιοι (2), τετρακισχιλίους (2), τετρακισχιλίων (1), τετρακόσια (2), τετρακοσίοις (1), τετρακοσίων (1), τετραπλοῦν (1), τρεῖς (43), τρία (8), τριάκοντα (11), τριακοσίων (2), τρισίν (6), τρισχίλιαι (1), τρίτη (3), τρίτῃ (13), τρίτην (3), τρίτης (2), τρίτον (16), τρίτος (7), τρίτου (3), τριῶν (11), χιλίας (1)

It would have been helpful, I think, if other quantifier words like πάς and ὅλος where included in this category as well.

Like Robinson and Logos, they strangely use the term “Definite Article.”

Their category “Particle” contains the fewest subcategories with only two: Conditional and Enclitic. I think this is a strange pair simply because the first of these labels is semantic (conditional) and the other is phonological. What is stranger is that Swanson’s “conditional particle” category consists of one word that occurs six times in the New Testament: εἴπερ.

The rest of Swanson’s particles are:

ἄγε (1), ἄν (166), ἄρα (49), ἆρα (2), Ἆρά (1), γάρ (1041), γε (25), δή (5), ἐάν (61), εἰ (426), εἴ (76), ἤ (342), ἡνίκα (2), ἤπερ (1), ἴδε (33), ἰδού (200), κάθου (1), κἄν (17), μέν (179), Μενοῦν (1), μενοῦνγε (3), μέντοι (8), μήν (1), μήτι (17), μήτιγε (1), νή (1), ὅταν (123), ὅτε (103), ὄφελον (3), ὄφελόν (1), πως (15), πῶς (103), τε (215), τοιγαροῦν (2), τοίνυν (3), ὡς (502), ὥς (2), ὡσεί (21)

What is strange about this list is that conditional conjunctions are only partially included here – and ironically, one of these, εἰ, is a proclitic, not enclitics. Also I do not understand why many of these words that are not clitics of any kind are included in this list, which only makes me want to ask “why?” a lot. At this point, a written description and explanation is needed.

Beyond this there isn’t much new. Swanson’s pronoun categories are quite basic and clear though they include demonstratives.


There are some striking differences and similarities between these categories. Friberg and Swanson are the most distinct. Friberg places certain words in significantly different places, while Swanson adds a number of new categories and subdivides “particles” rather strangely. Each database has its own idiosyncrasies and variations. Considering the differences seen thus far, I would suggest it wise to use multiple databases for morphological searching depending on what you are searching for. the results might be very different.

Previous Posts: Part I

Later Posts:

Part III: Conjunctions

Part IV: Verbs