Hilário de Sousa

Study notes

These are linguistic notes for myself: they are not fully referenced, and there are not many examples. If they help you, great. n.b.: I cannot guarentee the accuracy of these notes.

Abkhaz

I had a quick look at:

Alphabet/a little bit of phonology

Here are the Abkhaz alphabet organised phonologicaly. The inventory is that of the standard Abzhywa dialect (other dialects have more consonants). These are the consonants:

                                                                     
П п Т т Тә тә К к Кь кь Кә кә Ҟ ҟ Ҟь ҟь Ҟә ҟә
/pʼ/ /tʼ/ /tʷʼ/ [t͡pʼ] /kʼ/ /kʲʼ/ /kʷʼ/ [kʷʼ] /qʼ/ /qʲʼ/ /qʷʼ/ [qʷʼ]
Ҧ ҧ Ԥ ԥ Ҭ ҭ Ҭә ҭә Қ қ Қь қь Қә қә
/pʰ/ /tʰ/ /tʷʰ/ [t͡pʰ] /kʰ/ /kʲʰ/ /kʷʰ/ [kʷʰ]
Б б Д д Дә дә Г г Гь гь Гә гә
/b/ /d/ /dʷ/ [d͡b] /ɡ/ /ɡʲ/ /ɡʷ/ [ɡʷ]
М м Н н
/m/ /n/
Ф ф С с Ш ш Шь шь Шә шә Х х Хь хь Хә хә Ҳ ҳ Ҳә ҳә
/f/ /s/ /ʂ/ /ʂʲ/ [ʃ] /ʂʷ/ [ʃᶣ] /χ/ /χʲ/ /χʷ/ [χʷ] /ħ/ /ħʷ/ [ħᶣ]
В в З з Ж ж Жь жь Жә жә Ҕ ҕ Ӷ ӷ Ҕь ҕь Ӷь ӷь Ҕә ҕә Ӷә ӷә
/v/ /z/ /ʐ/ /ʐʲ/ [ʒ] /ʐʷ/ [ʒᶣ] /ʁ/ /ʁʲ/ /ʁʷ/ [ʁʷ]
Ҵ ҵ Ҵә ҵә Ҿ ҿ Ҷ ҷ
/tsʼ//tsʷʼ/ [tɕᶠʼ] /tʂʼ/ /tʂʲʼ/ [tʃʼ]
Ц ц Цә цә Ҽ ҽ Ч ч
/tsʰ//tsʷʰ/ [tɕᶠʰ] /tʂʰ/ /tʂʲʰ/ [tʃʰ]
Ӡ ӡ Ӡә ӡә Џ џ Џь џь
/dz/ /dzʷ/ [dʑᵛ] /dʐ/ /dʐʲ/ [dʒ]
Р р И и Ҩ ҩ У у
/r/ /j, jə, əj/ /jʷ/ [ɥ~ɥˤ] /w, wə, əw/
Л л
/l/

For the letter pairs <Ҧ Ԥ> /pʰ/ and <Ҕ Ӷ> /ʁ/, The ones with descenders <Ԥ Ӷ> are the modern standards, but the ones with hooks <Ҧ Ҕ> are still widely used. The ones with descenders <Ԥ Ӷ> are encoded in unicode since version 5.2 (i.e. later than the ones with hooks), and might not render correctly in some devices. (<Ꚋ Ӄ> also existed in earlier Cyrillic orthography of Abkhaz, for mordern <Ҭ Қ>.) There is also a (marginal) phoneme /ʔ/, but I don't know whether it is written or not.

There are two vowel phonemes: /ə/ /a/. There is also a long /aa/, which came from an earlier */ʕa/ or */aʕ/. There are the vowel symbols <Ы ы> for /ə/ and <А а> for /a/, and also <И и У у Е е О о> for other realisations of the vowel phonemes:

(In reality, things with the vowels are more complicated that this; let me read more about it / listen more first.) Stress is contrastive, but not indicated in the script.

Burmese

This section is primarily about the most common reditions of rimes in Burmese script, and their pronunciation in modern standard Burmese. (Please contact me if you see errors; I am still learning.) I am presenting them in two different ways that I find useful: arrangement based on dictionary order, and arrangement based on their phonology in spoken language. Unicode Burmese font is used here; the most commonly used font in Myanmar, Zawgyi, is encoded differently to Unicode. (On my computer at least,) Firefox renders Burmese script more accurantely than Chrome, so you might want to compare the rendition of Burmese script in different browsers.

Burmese script is follwed by MLC transcription, in italics. The MLC transcription is a transliteration of the written script; like many Brahmic scripts in Southeast Asia, there is considerable distance between the written script and the modern day pronunciation. You might want to have a look at, e.g., the Wikipedia pages on Burmese phonology and Burmese script.

Rimes in dictionary order

I read about the collation of Burmese script here:

Burmese script is a Brahmic script. Within each orthographic syllable, five types of script elements can be distinguished for the purpose of collation. These script elements largely correlate with the components in a phonological syllable. The collation hierarchy of the script elements is as follows:

For instance, in ကြောင့် kraung. 'because of', က k is the onset, ြ -r- is the onset diacritic, င် -ng is the coda, ော au is the vowel, and ့ . is the tone.

Onset

Basicaly follows the Brahmic script order. See the table here. In the table, read from left to right, and then top to bottom row by row. For the two varients of ny, ဉ precedes ည. Ignore the 'independent vowels' symbols for now.

Onset diacritic

The order is:

Coda

Basically any consonant symbol can be used to symbolise a coda by adding the ် a.sat diacritic to signify that it has no phonological vowel. For instance, က is ka., while က် is -k. They have the same order as the onset consonant symbols. Another way to symbolise a coda is using stacked consonant symbols: the normal consonant symbol is the coda, and the consonant symbol placed underneath is the onset of the following syllable. For instance, in မန္တလေး manta.lei: 'Mandaley' (the second city of Burma), န္တ n and t are stacked together; န n is the coda, and the တ t underneath it is the onset of the following syllable (in this case, /t/ becomes voiced /d/). The stacked consonants are collated in the same way as a normal consonant symbols.

Vowel

Some of the orthographic vowel symbols convey information more than just the phonological vowel. For their phonological values, see the rime table below. The following is the collation order of the orthographic vowel symbols; the vowel symbols are here placed around the zero onset symbol အ. The vowel symbol for a has two forms: the long form ါ and the short form ာ. The short form ာ is the default form; the long form ါ is used after ခ hk, ဂ g, ဒ d, ပ p, and ဝ w, to avoid them looking like other consonant symbols if they are followed by the short form ာ. The long form ါ is here accompanied by the onset ဝ w instead of the zero onset symbol အ. The vowel transcription is that of the orthographic vowel when it is not followed by a coda symbol.

Tone

The collation order is as follows:

Rime table in dictionary order

The following table presents the commonly encountered rimes in Burmese script. In each section the rimes of one or a pair of related vowel symbols are displayed. In the first one or two rows are rimes with no tone symbols, the following row are rimes with a ့ tone symbol, and the following row are rimes with a း tone symbol. (အော် (ဪ) au is put in its own row.) Each column are rimes with different coda symbols, starting with the first column with no coda symbol. For dictionary collation order, you need to read from top to bottom, and then from left to right column by column, because the coda symbol has higher priority than the vowel symbol in collation. The zero/ʔ onset symbol အ is used for the onset. Burmese script is followed by transcription in the same cell, and then its phonological form in the cell below. See also the notes following the table. (An online spreadsheet of this is also available here.)

a. အက် ak အစ် ac အတ် at အပ် ap
[ʔa̰] [ʔɛʔ] [ʔɪʔ] [ʔaʔ] [ʔaʔ]
အာ a အင် ang အဉ် any အည် any အန် an အမ် / *အံ am အယ် ai
[ʔà] [ʔɪ̀N] [ʔɪ̀N] [ʔɛ̀] / [ʔì] / [ʔè] [ʔàN] [ʔàN] [ʔɛ̀]
အင့် ang. အဉ့် any. အည့် any. အန့် an. အမ့် / *အံ့ am.
[ʔɪ̰N] [ʔɪ̰N] [ʔɛ̰] / [ʔḭ] / [ʔḛ] [ʔa̰N] [ʔa̰N]
အား a: အင်း ang: အဉ်း any: အည်း any: အန်း an: အမ်း / *အံး am:
[ʔá] [ʔɪ́N] [ʔɪ́N] [ʔɛ́] / [ʔí] / [ʔé] [ʔáN] [ʔáN]
အိ (ဣ) i. အိတ် it အိပ် ip
[ʔḭ] [ʔeɪʔ] [ʔeɪʔ]
အီ (ဤ) i အိန် in အိမ် / အိံ im
[ʔì] [ʔèɪN] [ʔèɪN]
အိန့် in. အိမ့် / အိံ့ im.
[ʔḛɪN] [ʔḛɪN]
အီး i: အိန်း in: အိမ်း / အိံး im:
[ʔí] [ʔéɪN] [ʔéɪN]
အု (ဥ) u. အုတ် ut အုပ် up
[ʔṵ] [ʔoʊʔ] [ʔoʊʔ]
အူ (ဦ) u အုန် un အုမ် / အုံ um
[ʔù] [ʔòʊN] [ʔòʊN]
အုန့် un. အုမ့် / အုံ့ um.
[ʔo̰ʊN] [ʔo̰ʊN]
အူး (ဦး) u: အုန်း un: အုမ်း / အုံး um:
[ʔú] [ʔóʊN] [ʔóʊN]
အေ (ဧ) e
[ʔè]
အေ့ e.
[ʔḛ]
အေး (ဧး) e:
[ʔé]
အဲ ai:
[ʔɛ́]
အဲ့ ai.
[ʔɛ̰]
အော (ဩ) au: အောက် auk အောင် aung
[ʔɔ́] [ʔaʊʔ] [ʔàʊN]
အော့ au. အောင့် aung.
[ʔɔ̰] [ʔa̰ʊN]
အောင်း aung:
[ʔáʊN]
အော် (ဪ) au
[ʔɔ̀]
*
အို ui အိုက် uik အိုင် uing
[ʔò] [ʔaɪʔ] [ʔàɪN]
အို့ ui. အိုင့် uing.
[ʔo̰] [ʔa̰ɪN]
အိုး ui: အိုင်း uing:
[ʔó] [ʔáɪN]

Some notes:

In addition, in cases with the consonant diacritic ွ -w-:

အွတ် wat အွပ် wap
[waʔ] / [ʔʊʔ] [waʔ] / [ʔʊʔ]
အွန် wan အွမ် wam
[wàN] / [ʔʊ̀N] [wàN] / [ʔʊ̀N]
အွန့် wan. အွမ့် wam.
[wa̰N] / [ʔʊ̰N] [wa̰N] / [ʔʊ̰N]
အွန်း wan: အွမ်း wam:
[wáN] / [ʔʊ́N] [wáN] / [ʔʊ́N]

In Upper Burma/Mandaley, the pronunciation is [waʔ] / [waN], while in Lower Burma/Yangon, the pronunciation is [ʊʔ] / [ʊN], and apparently this has merged into [ɪʔ] / [ɪN] for some people middle aged and younger. Both the Upper and Lower Burmese pronunciations of these are considered standard.

Rimes organised phonologically

The following are the rimes orgnised phonologically. The onset symbol used is the zero/glottal stop onset အ, or ဝ w when followed by the long form ါ a. Also read the section above for further explanations (e.g. how syllables can be reduced).

creaky tone . low tone high tone :
[a] a. အာ / ဝါ a အား / ဝါး a:
[i] အိ (ဣ) i. / အည့် any. အီ (ဤ) i / အည် any အီး i: / အည်း any:
[u] အု (ဥ) u. အူ (ဦ) u အူး (ဦး) u:
[e] အေ့ e. (/ အည့် any.) အေ (ဧ) e (/ အည် any) အေး (ဧး) e: (/ အည်း any:)
[ɛ] အဲ့ ai. / အည့် any. အယ် ai / အည် any အဲ ai: / အည်း any:
[o] အို့ ui. အို ui အိုး ui:
[ɔ] အော့ / ဝေါ့ au. အော် / ဝေါ် (ဪ) au အော / ဝေါ (ဩ) au:
[aʔ]/[aN] အတ် at / အပ် ap အန့် an. / အမ့် am. / အံ့ am. အန် an / အမ် am / အံ am အန်း an: / အမ်း am: / အံး am:
[ɪʔ]/[ɪN] အစ် ac အင့် ang. / အဉ့် any. အင် ang / အဉ် any အင်း ang: / အဉ်း any:
[ʊʔ]/[ʊN] အွတ် wat / အွပ် wap အွန့် wan. / အွမ့် wam. အွန် wan / အွမ် wam အွန်း wan: / အွမ်း wam:
[ɛʔ] အက် ak
[aɪʔ]/[aɪN] အိုက် uik အိုင့် uing. အိုင် uing အိုင်း uing:
[aʊʔ]/[aʊN] အောက် / ဝေါက် auk အောင့် / ဝေါင့် aung. အောင် / ဝေါင် aung အောင်း / ဝေါင်း aung:
[eɪʔ]/[eɪN] အိတ် it / အိပ် ip အိန့် in. / အိမ့် im. / အိံ့ im. အိန် in / အိမ် im/ အိံ im အိန်း in: / အိမ်း im: / အိံး im:
[oʊʔ]/[oʊN] အုတ် ut / အုပ် up အုန့် un. / အုမ့် um. / အုံ့ um. အုန် un / အုမ် um / အုံ um အုန်း un: / အုမ်း um: / အုံး um:

Georgian

I read through this:

I read parts of these:

Flagging and indexing of core gramamtical relations

In Georgian, the inventories of case and personal affixes that can be used to mark core grammatical relations is small. Nonetheless, the same markers are used in different ways, based on the verb class, and also the tense-aspect-mood of the verb for some verb classes. With these complex patterns, and some other morphosyntactic behaviours, different linguistics have been defining/arguing for different mappings between the arguments and the grammatical relations based on different morphosyntactic criteria.

Verb classes. The verbs are commonly divided into four classes. (Some linguists make finer distinctions.) I will simply call them Classes I, II, III, and IV. Prototypically:

There are many cases within each class that deviates from these transitivity prototypes. These verb classes are also characterised by other verbal morphologies, most of which are not discussed here.

Tense-aspect-mood. Each paradigm of tense-aspect-mood is traditionally called in English 'screeve' (from მწკრივი mts'k'rivi 'row'). The screeves can be divided into three series (the labels of the individual screeves often differ slightly in different discriptions:

Core cases. There are three core cases, they are usually called 'nominative', 'ergative', and 'dative'. Their forms vary depending on whether the noun/adjective stem they are attached to ends in a consonant or a vowel.

There are some class IV verbs of which the stimulus (the "object") is marked with a genitive case (noun: -is/-s, stems ending in -a or -e usually have -a/-e deleted and take -is; adjective: -i/-∅), instead of the usual nominative case. However, one line of thought is that since these genitive-marked nominals are not marked with a core case, nor cross-referenced, these nominals are oblique. These cases are not discussed further here.

There are three case marking patterns (-თვის -tvis below is a postposition that governs the genitive case):

Case pattern ↓ subject direct object indirect object
A ERG NOM DAT
B NOM DAT DAT
C DAT NOM GEN-tvis

These three case marking patterns are used with the following verb classes/screeves:

Verb ↓ | Case ↘ | TAM → Series I Series II Series III
Class I B A C
Class II B B B
Class III B A C
Class IV C C C

With case pattern C, there is also Harris (1981)'s opinion that pattern C is a derivative of pattern B, and in pattern C, the DAT-marked nominal is the indirect object, the NOM-marked nominal the subject, and the tvis nominal an oblique object. The derivation from the normal pattern B is as follow: i) the original subject underwent 'inversion', i.e. demotion to indirect object, and hence marked with DAT; ii) the original indirect object becomes an oblique, marked with -tvis; and iii) since there is now no subject, 'unaccusativity' is activated, i.e. the original direct object is promoted to subject, and hence this nominal is marked with NOM. Please see Harris (1981) for her arguments and exact wording. One fact that makes this account for pattern C attractive is that the case marking aligns with the cross-referencing: i) the NOM-marked nominal, traditionally regarded as the direct object, is now regarded as the subject, and it aligns with set I cross-referencing on the verb, which is usually used for subjects; ii) the DAT-marked nominal, traditionally regarded as the subject, is now regarded as the indirect object, and it aligns with the set II cross-referencing, which is usually used for objects; iii) the tvis-marked nominal, traditionally regarded as the indirect object, is now regarded as an oblique, and this correctly matches the fact that this tvis-marked nominal is not cross-referenced. This set of operations occur in TAM series III for verb classes I and III, and in verb class IV. (She also had to formulate that these do not occur twice for class IV verbs in series III.)

Nonetheless, the world is not perfect, and there is one trait that suggests that synchronically, in some ways speakers view the DAT-marked nominal in pattern C as the subject: for third person references, plural subject references can be cross-rerenced by a plural suffix, whereas objects are never cross-referenced as plural; in pattern C, when both arguments are third person, it is the DAT-nominal that can have its plurality cross-referenced by a plural suffix -თ -t, while the NOM-nominal cannot.

Harris' account of the DAT/NOM/-tvis nominal in pattern C being IO/SUB/oblique is at least diachronically correct, while the traditional account of the DAT/NOM/-tvis nominal being SUB/DO/IO is also not without reason. Here I will simply acknowledge that things don't line up nicely, and, when dealing with pattern C, simply call them the DAT-relation/NOM-relation/tvis-relation. (While for patterns A and B, I'll use the terms SUB/DO/IO as usual.)

Cross-referencing. There are two sets of verbal cross-referencing affixes in Georgian (an agreeing nominal does not need to occur): Set I and Set II. Set I is used for subjects. Set II is used for objects, with separate affixes for direct and indirect objects for third person. For case pattern C, Set I is used for the NOM-relation, and set II (indirect object) is used for the DAT-relation. In some verb forms, an auxiliary verb is used in place of a normal set I affix; this auxiliary verb is laregly identical in form as the copula, which itself contains set I affixes. An example is მ-ი-ყვარ-ხარ m-i-q'var-xar 'I love you', where 'I' is cross-referenced with a set II m-, and 'you' is cross-referenced by the auxiliary -xar, identical in form to the copular ხარ xar 'you (sg) are'.

Set I singular plural
1st person ვ- ვ-[...]-თ
v- v-[...]-t
2nd person ∅(/ხ-) ∅(/ხ-)[...]-თ
∅(/x-) ∅(/x-)[...]-t
3rd person -ს/-ა/-ო -ან/-ენ/-ნენ/-ეს
-s/-a/-o -an/-en/-nen/-es
Auxiliary (Set I) singular plural
1st person (ვ-)[...]-ვარ (ვ-)[...]-ვართ
(v-)[...]-var (v-)[...]-vart
2nd person -ხარ -ხართ
-xar -xart
3rd person -ა/-ს -ა/-ს
-a/-s -a/-s

The third person allomorphs are morphologically conditioned; it depends on the verb class and screeve. The second person ხ- x- is rare: a) in the present indicative copulas ხარ xar 'you(sg) are' and ხართ xart 'you(pl) are'; b) in the motion verb in the future and aorist series (-თ -t in the following indicates plural subjects): future მო-ხ-ვალ-(თ) mo-x-val(-t) 'you will come', conditional მო-ხ-ვიდ-ოდ-ი(-თ) mo-x-vid-od-i(-t), future subjunctive მო-ხ-ვიდ-ოდ-ე(-თ) mo-x-vid-od-e(-t); aorist მო-ხ-ვედ-ი(-თ) mo-x-ved-i(-t) 'you came', aorist subjunctive (optative) მო-ხ-ვიდ-ე(-თ) mo-x-vid-e(-t). (The preverb მო- mo- can be substituted by other preverbs, e.g. მო-ხ-ვალ mi-x-val 'you will go'.)

Set II singular plural
1st person მ- გვ-
m- gv-
2nd person გ- გ-[...]-თ
g- g-[...]-t
3rd person direct object ∅- ∅-
∅- ∅-
3rd person indirect object ∅-/ს-/ჰ- ∅-/ს-/ჰ-[...](-თ)
∅-/s-/h- ∅-/s-/h-[...](-t)

As for the set II third person indirect object affixes: a) ს- s- in front of ც- წ- ძ- ჩ- ჭ- ჯ- თ- ტ- დ- | ts- ts'- dz- tš- tš'- dž- t- t'- d- (all the coronal plosives and affricates); b) ჰ- h- in front of ქ- კ- გ- ყ- | k- k'- g- q'- (all the dorsal plosives) or პ- p'-; and c) ∅ elsewhere. (One line of thinking is that) third person objects never take the plural suffix -თ -t; nonetheless, in case marking pattern C, the DAT-relation is now viewed as the subject, and the corresponding set II third person affix can take the plural suffix -თ -t if the NOM-relation is also third person.

A verb must have a set I affix. If the relation cross-referenced by a set II affix is empty, it is cross-referenced as 3SG direct object, i.e. ∅, by default. These prefixes and suffixes occur at the outer-most periphery of the verb, with the exception that preverbs can exist in front of a cross-reference prefix.

Combining cross-referencing affixes. There is a constraint that at most one non-zero prefix and one non-zero suffix can occur (the internal structure of the copula-like auxiliaries does not count). When more than one is logically needed in the prefix or suffix slot, there is a hierarchy as to which one is kept and which one is deleted (in the list below, > signifies 'has precedence over'):

For prefixes, see below for what happens when there is a 2SG direct object and 3SG indirect object. For the suffixes, it can be reworded like this: any plural suffix has precedence over the plural suffix -თ -t or a singular suffix.

These deletion rules can create two homonym sets, e.g. გმალავთ gmalavt 'we hide you(sg)', 'I hide you(pl)', 's/he hides you(pl)', or 'we hid you(pl)'; გმალავენ gmalaven 'they hide you(sg)' or 'they hide you(pl)'. In the table below, მალ mal is the verb root 'hide', and -ავ -av is a 'thematic suffix' (which is used primarily in the present/future series; in this case the verbs are in present indicative). Set I indexes the subject, and Set II the object, so, for instance, Set I 2SG and Set II 1sg მმალავ mmalav means 'You(sg) hide me'.

Set I → | Set II ↓ 1SG 2SG 3SG 1PL 2PL 3PL
1sg მ-მალ-ავ მ-მალ-ავ-ს მ-მალ-ავ-თ მ-მალ-ავ-ენ
m-mal-av m-mal-av-s m-mal-av-t m-mal-av-en
1pl გვ-მალ-ავ გვ-მალ-ავ-ს გვ-მალ-ავ-თ გვ-მალ-ავ-ენ
gv-mal-av gv-mal-av-s gv-mal-av-t gv-mal-av-en
2sg ვ-გ-მალ-ავ გ-მალ-ავ-ს ვ-გ-მალ-ავ-თ გ-მალ-ავ-ენ
v-g-mal-av g-mal-av-s v-g-mal-av-t g-mal-av-en
2pl ვ-გ-მალ-ავ-თ გ-მალ-ავ-ს-თ ვ-გ-მალ-ავ-თ-თ გ-მალ-ავ-ენ-თ
v-g-mal-av-t g-mal-av-s-t v-g-mal-av-t-t g-mal-av-en-t
3 ვ-მალ-ავ მალ-ავ მალ-ავ-ს ვ-მალ-ავ-თ მალ-ავ-თ მალ-ავ-ენ
v-mal-av mal-av mal-av-s v-mal-av-t mal-av-t mal-av-en

Two more comments need to be made. Reflexivation is indicated by the noun თავი tavi 'head', sometimes preceded by a possessive pronoun, e.g. ჩემი თავი tšemi tavi 'myself' (3SG reflexive is თავისი თავი tavisi tavi). This reflexive nominal is cross-referenced as third person (and hence the combinations of first person–first person and second person–second person do not feature in the table above). For instance, ჩემს თავს ვ-ა-ქ-ებ tšem-s tav-s v-a-k-eb (my-DAT head-DAT 1SG-NV-praise-TS) 'I praise myself'. Often, subjective version is used (ი- i- in the following example), in which case the possessive phrase is not used, e.g. თავს მო-ვ-ი-კლ-ავ tav-s mo-v-i-k'l-av (head-DAT PV-1-SV-kill-TS) 'I shall kill myself'.

We have seen above that there cannot be more than one non-zero cross-reference prefix. When there are a first or second person direct object and an indirect object, a more drastic approach is used than simply deleting one of the object prefixes: the direct object nominal is changed to თავი tavi 'head' preceded by a first or second person possessive adjective, and it is cross-referenced on the verb as third person, i.e. zero. For instance, შენ-ს თავ-ს მ-ა-ძლ-ევ-ენ šen-s tav-s m-a-dzl-ev-en (your-DAT head-DAT 1SG-NV-give-TS-3PL 'they give you(r head) to me'.