This week, we worked with Harsha in the Linguistics seminar room, Pearson 115. We examined the number system and collected much new vocabulary, as well as opening up a wide range of possible exploration into the morphology (and perhaps historical change) of the number system.

Bengali appears to have a base-10 counting system. We elicited numbers from 1-20, then 30, 40, 50, ..., 90, 100. Harsha then prompted us to go further into the individual numbers within each block of ten, hinting that they followed different pattern. Below are the cardinal numbers from 1-49 - the full set of data (including 91-95 and 97-99, which Harsha was less certain of) can be found in a separate sheet on the dictionary, as there was an enormous amount.

Note that Mollie filled out the new dictionary sheet with her own notes, while the data below is based on Sam's notes; thus, there may be discrepancies in transcription, though all of Sam's transcribed lexical items are also listed in the dictionary in the "disputations" column. All discrepancies will be later solved through clarification with Harsha.

     0        10        20       30        40    
0             dɔʃ       kuri     tiriʃ     colliʃ
1    æk       ægaro     ekuʃ     ektiriʃ   ækalliʃ
2    dui1      baro      baiʃ     bottiriʃ  bialliʃ
3    tin      tæro      teiʃ     tettiriʃ  tiralliʃ/titalliʃ3
4    car      coddo     cobbiʃ   coutiriʃ  cualliʃ
5    pãc      ponero    pociʃ    pɔetiriʃ  pɔncalliʃ
6    chɔi     ʃolo      cabbiʃ   chɔttiriʃ chialliʃ
7    ʃat      ʃatero    ʃataʃ    ʃaitiriʃ  ʃatalliʃ
8    aʈh      athero2    aʈhaʃ    attiriʃ   aʈalliʃ
9    noi      oniʃ      untiriʃ  unocholiʃ unoponcaʃ

1. Harsha provided ba- or b[V]- as an alternative prefix for "two" and noted it was similar to English.   
2. Sam noted 18 as athero and 38 as attiriʃ, but 8, 28, 48, and 98 with retroflex /ʈ/. Mollie noted all as dental. Clarification with Harsha is needed here - though, as said above, there are discrepancies, a switch between dental and retroflex stops is unexpected.
3. Harsha wasn't certain which sounded right, and was unsure which of these he would use.

Bengali has separate monomorphemic words for all numerals from 1-10; we did not elicit any word for "zero" in this session. Later numbers appear to be combinations of a morpheme indicating the tens place and another indicating the ones place - all number in the 30s, for example, included tiriʃ. However, the exact nature and interactions of these morphemes is not consistent across groups of 10, as seen in the chart above.

A notable example not present in the chart is 99. Although many of the -9 numbers we elicited were of the form un(o)-[following decade], which Harsha confirmed was the norm, 99 breaks the pattern using instead niranobboi. *unækʃo was explicitly unacceptable.

Harsha described the biʃ morpheme seen in the 20s as another word for 20, though it is used only in counting. He drew a parallel between this and what he called the "root" form of some numbers, including /ek/, the "root form" of 1. He mentioned that /ek/ is "one" in Hindi, and it shows up in the ones-place morpheme (the first half) of many later numbers (see 21 and 31, though notably not 11, 41, or 91).

-----

We also looked further into a distinction that had been briefly mentioned and touched upon in earlier sessions: that between dui, du and duto as words for "two". Harsha gave the first when listing numbers, but indicated that the second and third are used when counting things.

Between du and duto, there appears to be a distinction of specificity or distribution when used to pluralize a noun. Harsha originally explained the difference as "du botol" being "two bottles" and "duto botol" as being "two specific bottles". He also said that he would use the former when buying bottles of milk from the milkman, but the latter when he was referring to two bottles he had lost. Harsha initially explained it as similar to the difference between "fish" (as a plural) and "fishes". He then mentioned that, when describing tables, he would use "du tebel" to mean any two of some table kind, while "duto tebel" would refer to any two tables, regardless of kind.

This prompted the drink scenario, as follows, which more clearly indicates the distinction between the two words:
   If you're at a party and there is a variety of drinks available, and you ask your friend to go and get you "du" drinks, would they respond differently than if you asked them to get you "duto" drinks?
The answer here was yes, they would:
   --- With du, they would either ask "what drink?" or return with two of the same drink.
   --- With duto, they would either ask "what drinks?" or return with two different drinks. 

Harsha indicated that one would use tin for 3 in the same situation as du, and tinte in the same situation as duto. He drew a parallel between the -/to/ and -/te/ endings here and the -/ta/ suffix on nouns. However, given the drink scenario with numbers higher than three (for instance, carte for four), Harsha said that one would likely get four of the same drink, simply because it's unlikely one would ask for four different drinks and so the request would not be interpreted as such.

----- 

We then examined the ordinal numbers from 1-10:

     1. In this case, rather than a single (adjective?) word for "first", Harsha used the phrase lain-er protom-e "in the first position of the line"; this is similar to the protoma used to denote the first day of the month.
     2. ditio   ---   Harsha notes here that one could not say "lainer ditiote" in the same way that one does with protom. This is of note because protome appears to consist of protom plus the locative -/e/ suffix; here, when giving a hypothetical combination of ditio with that suffix, there appears to be an epenthetic /t/ added.
     3. tritio
     4. coturtho
     5. pɔncom
     6. ʃɔʃʈom   ---   Harsha was unsure of this word; given here is his best guess.
     7. ʃɔʈom    ---   Harsha was unsure of this word; given here is his best guess.
     8. ɔʃʈom
     9. nɔbom
     10. dɔʃom

(Note that everywhere above where there is a retroflex /ʈ/, based on Sam's notes, Mollie recorded a dental. Clarification with Harsha is needed in these cases.  Further investigation into days of the month can be found here and is planned for our next group meeting.)

Harsha mentioned that after 10 there were no single words and he would use a construction such as dɔʃ jon-er phecon-e "behind ten people" for "eleventh", and so on with ægaro jon-er phecon-e "behind eleven people" for "twelfth", etc.

For the last person in a line, one could also use lain-er ʃeʃe "at the end of the line". 

There is no single word translatable as "penultimate", but in that case Harsha would use æk jon-er ʃamn-e "in front of one person", though he notes that this could also be interpreted as meaning "second in line". The ambiguity between "behind" and "in front of" when referring to lines of people or things is also noted in English, and a more semantically-focused small group session may wish to explore this further.

-----

There appears to be a good deal of phonological variation between groups of ten, and Harsha speculated that it may be due to historical change, which we agreed with; a future elicitation session to acquire numbers from 51-89 would be ideal to get a full picture and possibly find patterns relating to the notable features following. 
Look further into:
     ---   Cardinal numbers from 51-89, 101+
     ---   Vowel changes in tens-place morphemes 
     ---   Extremely high presence of geminate consonants
      ---   Use of "root" or "alternate" forms of numbers for counting (e.g. biʃ)
      ---   Specific behavior of du and duto in different contexts, additional differences between numbers as elicited and numbers used when pluralizing (note that the cardinal number 2 is not du but dui)
      ---   What's up with that epenthetic /t/ in the hypothetical-but-incorrect "ditiote"? Have we seen -/e/ added to other /o/-final words with similar results?
      ---   Might be nice to ask Paroma about words Harsha was unsure of, including "sixth" and the 91-99 set.