Feelyat: A Corpus-Based Analysis

[Emily]

For ages now I have been saying I want to do a corpus-based analysis of the language used in the classic Kids in the Hall sketch Feelyat and I finally decided now's the time!!

For starters, here's my best guess at a transcription, followed by the apparent (approximate) meanings

#	Speaker	Utterance	Tenative translation
1	ANNOUNCER:	ei! ɛs gəˈdurə ˈilda ˈfiljat?	Hey! Who's ready for Feelyat! (or similar slogan)
2	AUDIENCE:	[?? vɔs?] du ˈfiljat!	(second part of slogan; first word could be hɔs, ɔs, or possibly wʌts; if the latter, this would likely mean "What do you Feelyat?")
3	Text:	feelyat!	feelyat!
4	ANNOUNCER:	i ˈvala di ˈhostu, ˈvalda ˈgrɛbnɛs!	And here's the host, Valda Grebnes!
5	HOST:	ˈprivu! ˈprivu! ˈprivu.	Thank you! Thank you! Thank you.
6	HOST:	ˈunas dəlˈvitəl gɚ ˈʃakə ˈfilijat, ˈkrejo pɚ ʊnˈdʒɪtsalou, ˈvænjə smɪts!	(exact meaning unclear, but clearly introducing the hostess, Vænjə Smits; second word might be vəlˈvitəl)
7	HOSTESS:	aˈlou, ˈvalda!	Hello, Valda!
8	HOST:	aˈlou, ˈvænjə.	Hello, Vænjə.
9	HOST:	torks də flɚ?	(unclear)
10	HOSTESS:	jɔ jɔ!	Yes, yes!
11	HOSTESS:	hat ˈvalda, sou ɛs dɚ ˈnɛdɚlændɚ fʊt ˈkwajɚ!	But Valda, so is the Dutch Foot Choir!
12	HOST:	di ˈnɛdɚlændɚ fʊt ˈkwajɚ!	The Dutch Foot Choir!
13	HOST:	ni lats ˈfilijat!	Now let's Feelyat!
14	HOST:	suva, kan ˈsaga val ˈda.u kanˈtɛstas?	Suva, can you introduce the contestants?
15	ANNOUNCER:	ˈkrevu, valda!	Of course, Valda!
16	ANNOUNCER:	jir, di tʃampiˈɔn van ˈantvɚp ˈbeldʒʊm, hans ˈbodɪs!	Here (is) the champion, from Antwerp, Belgium, Hans Bodis! (final consonant in last name unclear)
17	ANNOUNCER:	ʊn ˈdate prɛmiˈɛr tʃaˈlɛndʒɛr, ˈoutən ˈhanaboum ə ˈʃvedən, gəˈhɛljo ˈbonɪk, ˈrɪki srajʊnˈkɔvsɛn!	And his first challenger, out of Hannebom(?), Sweden, (unclear) Ricky Sraiunkovsen!
18	ANNOUNCER:	ɔn dɛr sɛˈkundɚ tʃaˈlɛndʒɛr vɔn tɔˈrɔnto ˈkænəda, daˈrɪl ˈvasdɪk!	And the second challenger, from Toronto, Canada, Darill Vasdik!
19	ANNOUNCER:	ʊntan, daˈrɪl! ha ha ha ha!	Welcome, Darill! Ha ha ha ha!
20	HOST:	hans, hɔrz də ˈmanzi ˈfilijat?	Hans, are you excited for Feelyat? (or are you ready for Feelyat or something similar)
21	OLD MAN:	nei.	No.
22	HOST:	hɔrz də ˈmanzi ˈfilijat?	Are you excited for Feelyat?
23	OLD MAN:	nei.	No.
24	HOST:	ˈfilijat. hɔrz də ˈmanzi?	Feelyat. Are you excited?
25	OLD MAN:	ˈfiljat!	Feelyat!
26	HOST:	ni lats ˈfilijat!	Now let's (play) Feelyat!
27	HOST:	ˈsuva, wat ɪn dai ˈabdʒɛkt?	Suva, what is the (or "his") object?
28	ANNOUNCER:	dai ˈabdʒɛkt ɪz in ˈpainæpəl.	The (or "His") object is a pinaeapple.
29	HOST:	ʃː! trein seˈgand mɪt di ˈnɛdɚlændɚ fʊt ˈkwajɚ!	Sh! Twenty seconds with the Dutch Foot Choir!
30	OLD MAN:	ɪʃt ain prak? nei. ˈdɛrti faʃ? nei. ɪʃt ain ˈhɛʃta? ɛ! ˈhɛʃta? nei, nei, nei, nei, nei, nei, nei. ˈhɛʃta biʃ—	Is it a (unclear)? No. Is it a (unclear)? Hmm! (unclear)? No, no, no, no, no, no, no. (unclear)—
31	Text:	NEWSIN BULLETUNIC	News Bulletin
32	NEWS ANCHOR:	ni ˈmɔntuʃ trəˈdʒuki.	(something on the order of We interrupt this broadcast or Breaking news; first word is likely "Now")
33	NEWS ANCHOR:	dɪˈzastʊr zʊn rain.	Disaster on the Rhine.
34	Text:	Disastür zün Rhine	Disaster on the Rhine.
35	NEWS ANCHOR:	ˈviʃən ut ˈælən ˈata ˈkrambəltɔk, ˈvradə ut ˈtɔpɔt, blak ˈtunə ˈɛbuʃ frɛm.	(unclear)
36	NEWS ANCHOR:	rəˈpitən: ˈviʃən ut ˈælən ˈata ˈkrambəltɔk, ˈvradə ut ˈtɔpɔt, blak ˈtunə ˈɛbuʃ frɛm.	Repeat: (unclear)
37	NEWS ANCHOR:	ni ˈpunək poi ˈfiljat!	Now back to Feelyat!
38	HOST:	ˈrɪki, wʌts du ˈfili?	Ricky, what do you feel?
39	SMALL BOY:	ˈɛsən ist ain pip?	Is it a (unclear, possibly "pipe" or "flute")
40	HOST:	nei, ain ˈklakən.	No, a clock.
41	AUDIENCE:	ɔː!	Aww!
42	HOST:	ˈvænjə?	Vænjə?
43	HOST:	dəˈrɪl ˈvasdɪk, wʌts du ˈfili?	Darill Vasdik, what do you feel?
44	WINNER:	ˈarsən dust ain mit?	Is this a meat?
45	HOST:	ˈablia! ˈablia, ˈablia.	Wait! Wait, wait. or: Quiet! Quiet, quiet.
46	HOST:	un spəˈsɪfɪka, wʊt ˈkundɚ mit?	Specifically, what kind of meat?
47	WINNER:	ˈɛsən ɪs ˈbiftɛk?	Is it (a) steak?
48	HOST:	ˈfilijat!	Feelyat!
49	HOST:	(inaudible over audience, but whatever it is he says it twice)	(unclear, probably general congratulations)
50	HOST:	ˈsuva, jənəlˈtuzən dʊlˈdaʃ daˈrɪl?	Suva, what has Darill won?
51	ANNOUNCER:	daˈrɪl vosˈpasɛn in viɛtnaˈmiz pɔt ˈbɛli pɪg!	Darill has won (or Darill will be taking home, etc.) a Vietnamese pot-bellied pig!
52	HOST:	ˈɪldiəmeo ˈfojalo ˈplatɪk ˈfilijat.	(unclear except for Feelyat)
53	HOST:	ˈpai pai! ˈpai pai! ˈpai pai.	Bye bye! Bye bye! Bye bye.

Right off the bat we can see this is almost certainly a Germanic language, with clear many clear cognates to English vocabulary; this supposition is supported by contextual clues such as the characters coming from different predominantly Germanic-speaking countries.

Different aspects of this language will be examined in individual posts. Feel free to contribute!

[Emily]

PHONOLOGY
Consonants
The consonant inventory is very similar to English, with no definite non-English consonants present:

Code: Select all

p b t d k g
f v s z ʃ h
tʃ dʒ (ts)
m n
j l r w

Of these, only the affricates are less than certain. /dʒ/ is almost definitely a phoneme, given the total absence of /ʒ/ in the corpus. Though both /t/ and /ʃ/ are distinct phonemes, it is likely that /tʃ/ is a separate phoneme as well, as it is found word-initially, an environment where a stop cannot otherwise be followed by a fricative (see below). Though it's possible /ts/ could be a distinct phoneme, it's equally possible that it is a simple consonant cluster. The rhotic /r/ seems to be in free variation between [ɾ] and [ɻ]; it is transcribed /r/ for convenience. Finally, there is no instance of a clear /ŋ/ phoneme in the corpus; the only instance of /n/ followed by a velar is in the name "Sraiunkovsen" (17), and it's difficult to hear whether the speaker is articulating [n] or [ŋ].

Vowels
The vowels present in the corpus are /a æ ɛ e ə ɚ ɪ i ɔ o ʌ ʊ u/. Of these, /ʌ/ is possibly an allophone of /a/, as it only appears in a word (/wʌts/, 38/43) that seems to alternate with /wat/ (27) (but see also /wʊt/, 46). It is likely that most instances of [ɚ] are analyzable as the realization of a sequence /ər/, but the former appears in stressed syllables as well (flɚ, 9), which /ə/ never does. Vowels seem to be able to be reduced to /ə/ in unstressed syllables, though this process does not seem as extensive as in English.

The diphthongs present in the corpus are /ai ei oi ou/. Some words transcribed with a vowel followed by /j/ may be interpreted as pure diphthongs as well. In the one appearance of the sequence /au/ in the corpus (dau, 14), the vowels are clearly distinguished as two separate syllables.

Stress
Determining the overall stress pattern of the language is necessarily hampered by the fact that word boundaries are sometimes unclear: when the announcer introduces the host, does he do so with a single word /iˈvala/ carrying penultimate stress, or with an unstressed word /i/ followed by the word /ˈvala/ carrying initial stress? I have transcribed it with the latter, but it could just as easily be the former. Nevertheless, the majority of multisyllabic words seem to have initial stress, with the exceptions mostly found in names (Darril, Toronto) and obvious borrowings (champion, challenger) — though see /trəˈdʒuki/ (32), /vosˈpasɛn/ (51).

Phonotactics
The languages has a wide range of syllable types: V, VC, CV, CVC, and so on, up to an apparent maximal syllable type of CCVCC (though the longest syllable in the corpus is CVCCC). There does not seem to be any restriction on what consonants may begin a syllable; based on related languages, it is likely that a syllable cannot end in /h/ or begin with /ŋ/ (if such a phoneme exists), and whether a syllable can end in /w/ or /j/ depends on how one analyzes diphthongs (see above). Outside of personal or place names, attested consonant clusters are /bl fl fr kl fr kw pl pr sp tr vr/, and attested coda clusters are /kt nd rz st ts ʃt rks/. Including names adds /gr sm sr ʃv/ to the onsets and /ns nt/ to the codas; these may or may not reflect the actual phonotactics of the language, as the bearers of these names come from at least three different countries, all of which have different national languages (none of which are the language under discussion). Still, it seems likely that most of these clusters are legal in the language itself. Though a larger corpus is needed to confirm the language's phonotactics, the following speculative list of legal consonant clusters seems reasonable based on the data and related languages:

• Onsets:
  • /t,d/ + /r/
  • /p,b,k,g,f,v,s,z/ + /r,l/
  • /s/ (and possibly /ʃ/) + /p,t,k,m,n,l,r/
  • /k,g/ + /w/
• Codas:
  • any voiced consonant + /z/
  • any unvoiced consonant + /s/
  • nasal + homorganic stop
  • /l,r/ + any stop
  • /s,ʃ/ + /p,t,k/
  • /p,k/ + /t/
  • any of the above + /s,z/ (as voicing dictates)

Initial clusters of the type /s/ + /p,t,k/ + /r/ (as well as /spl/ and /skw/) would be consistent with other Germanic languages, but do not appear in the corpus.

The lax vowels /ɛ ɪ ɔ ʊ/ (and, if indeed it is a phoneme, /ʌ/) do not appear word-finally except in certain interjections such as /jɔ/ "yes" (10) and the apparent filler word /ɛ/ (30). There are no words in the corpus in which a diphthong precedes a consonant cluster.

Individual and dialect variation
The host pronounces the word "Feelyat" as /filijat/ (see 6, 13, etc.) while the other speakers consistently say /filjat/ (1, 2, 25, 37); it is likely that many speakers pronounce a /j/ as /i/ or /ij/ following a consonant.

The elderly man pronounces a certain word /iʃt/ (30) while the young boy pronounces the same word /ist/ (39); the man may speak a dialect that pronounces standard /s/ before other consonants as /ʃ/ (commonly seen in High German varieties). If this is the case, his [hɛʃta] (30) may represent an underlying /hɛsta/ in the standard language.

Orthography
Only three lines of text are present in the corpus, and only two match up to spoken utterances: "feelyat" (3) is the spelling of /fil(i)jat/, the name of the show, and "Disastür zün Rhine" (34) appears right after the news anchor announces /dɪˈzastʊr zʊn rain/ (33). The spelling "ee" for /i/ and "y" for /j/ are unexpected given the provenance of the language, and likely reflect English influence; the same may also be surmised for the spelling "Rhine", though why this spelling would be taken from English rather than a continental language is unclear. In the headline, we can see two spellings for /z/: <s> in "Disastür", and "z" in "zün". We can make a reasonable guess that "s" represents /z/ intervocalically, and presumably an /s/ in this position would be spelled "ss". The symbol "ü" consistently represents /ʊ/. Reliable analysis of "NEWSIN BULLETUNIC" (31) is not possible without information as to how it is pronounced.

[Vijay]

I thought #4 was something like [tʰɹ̥ej buw], and I think I've seen it transcribed as <tray boo>, in which case it's presumably from French très bon or perhaps even très bien. Same in #15.

[Emily]

I thought #4 was something like [tʰɹ̥ej buw], and I think I've seen it transcribed as <tray boo>, in which case it's presumably from French très bon or perhaps even très bien. Same in #15.

Damn that was fast, you beat my edits lol

When I was transcribing it, I had trouble deciding between /krivu/ and /privu/, but listening again (and watching his lips) it could just as easily be /trivu/. And a /b/ sounds just as likely as a /v/. But the vowel sounds very clearly /i/ to me rather than /e(j)/, and I'm not really sure why he would be saying "very good" or "very well" in that line. Although damned if I can make a connection between "privu" and any word in the real world. It certainly makes more semantic sense in #15, but to me the announcer much more clearly has an initial velar (though /k/ vs /g/ was hard for me to distinguish) and a /v/ rather than a /b/. (It's hard to make exact judgments on these, of course; I should be at least pulling out my DVD copies and do this off of there instead of YouTube!)

[Vijay]

But the vowel sounds very clearly /i/ to me rather than /e(j)/, and I'm not really sure why he would be saying "very good" or "very well" in that line.

That's quite possible. I think maybe he meant something like 'all right, all right!'.

[Emily]

GRAMMAR
Articles
It's impossible to construct a complete article system from this short corpus. However, the language clearly has both definite and indefinite articles, and given the variation in forms the articles are likely to agree with the noun in gender and probably case and number as well.

Definite article
Multiple forms of definite articles are present in the corpus, but analysis is difficult. The announcer uses the article /di/ with the word /tʃampiɔn/¹⁶ and /dɛr/ with /tʃalɛndʒɛr/,¹⁸ despite both of these phrases clearly appearing in the same case, both clearly being singular, and both presumably in masculine gender. It's possible some morphological quirk overrides the natural gender for one of these words, much like the diminutive suffix -chen renders German Mädchen "girl" a neuter noun, though it isn't clear why this would be the case.

Another tricky example is the word /kwajɚ/. The hostess uses the article /dɚ/¹¹ (likely an unstressed variant of the announcer's /dɛr/; the announcer tends to enunciate more distinctly than the other speakers, and seems to avoid /ə/ where possible), while the host uses the article /di/,¹² both in positions that appear to be nominative. Two possible solutions to this problem suggest themselves:

1. An oblique case may be used when no predicate is present. A similar phenomenon occurs for pronouns in conversational English and (I believe) in French, and it may very well be the case that this rule applies to other case-marked parts of speech in this language as well.
2. Alternatively, verbless introductions ("The Dutch Foot Choir, ladies and gentlemen!") may by convention use the accusative case, filling in an implied "May I present to you" or something similar.

In both of these scenarios, the nominative form would be the hostess's /dɚ/, while the host's /di/ would be an oblique case of some kind. (The announcer's introductions of the the contestants in 16-18 would be affected by either of these explanations as well, though neither resolves the questions raised by /tʃampiɔn/ vs. /tʃalɛndʒɛr/. Whether the scenarios affect the introduction of the host in 4 depends on whether that sentence has a verb or not.) It is further borne out by a later use of /di/ for the same noun when it is the object of a preposition,²⁹ where Germanic languages never to my knowledge use the nominative case.

There are also some appearances of ambiguous words that may or may not be definite articles. Firstly, the host uses the phrase /da.u kantɛstas/.¹⁴ This may be a plural article (if so, it is the only one decipherable in the corpus), it may be a contraction of an article and some other word, or it may not be an article at all.

Secondly, the announcer introduces the first challenger with /date prɛmiɛr tʃalɛndʒɛr/¹⁷ but the second with /dɛr sɛkundɚ tʃalɛndʒɛr/.¹⁸ It's possible that /date/ is an article (it certainly resembles dat found in many West Germanic varieties outside of the High German areas), but, given the completely different form found in nearly the exact same context in the immediately following sentence, it seems more likely that /date/ is some other determiner instead; I have glossed it above as "his" (though this is not without its own problems).

The word /abdʒɛkt/ is preceded twice^27,28 by the word /dai/, both in positions where we would expect a nominative. It's possible this may mean "his", but it's equally possible that it's another singular definite article.

Lastly, the word zün^33,34 may be a contraction of a preposition and a definite article, presumably ending in -n. However, it may be a simple preposition used without an article at all, in an example of "headlinese".

To summarize, the only clearly nominative definite articles are /dɛr ~ dɚ/¹¹ and (if it is an article at all!) /dai/.^27,28 The form /di/ serves at least in part as an oblique or prepositional declension of /dɛr ~ dɚ/, though the two forms also contrast in at least one grammatical environment.^16,18 And there is a possible plural definite article, /da.u/;¹⁴ if this is indeed a plural article, it is likely accusative.

Indefinite article
By contrast, only two forms are attested for the indefinite article, with almost all attestations appearing in nominative positions: /in/ is used with /painæpəl/,²⁸ and /ain/ is used with the nouns /prak/,³⁰ /pip/,³⁹ /klakən/,⁴⁰ and /mit/.⁴⁴ The article /in/ is used for the noun /pɪg/⁵¹ in the accusative case, but this word is not attested in the nominative. No indefinite plural is attested.

[Emily]

Verbs
Very few words in the corpus are clearly interpretable as verbs: the news anchor's /rəpitən/³⁶ is likely a gerund, and the host's questions /torks də flɚ/⁹ and /hɔrz də manzi filijat?/^20,22,24 may begin with verb stems /tork-/ and /hor-/ with a 2sg ending /s~z/ (likely /-s/ after voiceless consonants and /-z/ otherwise). The word /fili/ "feel" appears in the question /wʌts du fili/, and is likely an infinitive (its lack of a final -en calls into doubt the possibility that /rəpitən/ is an infinitive. Beyond these examples, the only clearly reconstructable verbs are all 3sg conjugations of to be.

Even this poses problems, given the wide variety of likely forms: it appears as /ɛs/,^1,11 /ɪn/,²⁷ /ɪz/,²⁸ /ɪʃt/,³⁰ /ist/,³⁹ /arsən/,⁴⁴ and /ɪs/.⁴⁷ I am unable to bring all these different forms into a coherent system, but I can get close, especially if we entertain the possibility that the language may have two copulas similar to Old English or Modern Spanish.

• The form /ist/ may represent more permanent characteristics, like Spanish es. The form /ɪʃt/ has already been described as a likely dialect variant, and /ɪs/ is probably a form of this word as well. It may be the case that /ɪz/ is part of this construction too, though the nature of the variation is unclear.
• The form /ɛs/ may represent temporary states, like Spanish estar.
• The form /arsən/ is likely a subjunctive: /arsən dust ain mit/ = "could this be a meat?"

This leaves only /ɪn/ unaccounted for. Unfortunately, I simply cannot explain its form and why it different from one of the other forms we would expect in its place.

Pronouns
A handful of pronouns are reconstructible. The 2sg pronoun /du/ appears multiple times;^2,38,43 the host's repeated /də/ after /torks/⁹ and /horz/^20,22,24 is probably an unstressed variant. The word /ɛsən/ appears twice, both times glossed in English as "it".^39,47 The word /dust/ is clearly a pronoun, and likely translates as "this".

The word for "what" appears in multiple forms: /wat/ as a nominative,²⁷ /wʌts/ as an accusative,^38,43 and /wʊt/ in an adjectival form "(what kind of meat") in a question;⁴⁶ this last appearance could be nominative ("what kind of meat is it") or accusative ("what kind of meat do you feel"). As noted in a previous post, it's possible that [ʌ] is an allophone of /a/, though it may be a phoneme in its own right.

[bradrn]

I love it when people start with an utterly ridiculous premise and treat it completely seriously. Bravo! (Or, should I say: Feelyat!)

[alice]

I love it when people start with an utterly ridiculous premise and treat it completely seriously.

Remember Bobbity-bob and his utterly ridiculous declarations of his greatness?

[bradrn]

I love it when people start with an utterly ridiculous premise and treat it completely seriously.

Remember Bobbity-bob and his utterly ridiculous declarations of his greatness?

Oh, yes, how could I forget? That wasn’t the same though: Bob started with pretty reasonable premises, he was just really terrible at communicating them to us.

(Though really I’d just prefer to forget the whole unfortunate episode. And preferably retract the horrible things I said to him, though there’s not much I can do about that now.)

[Emily]

Verbs
Very few words in the corpus are clearly interpretable as verbs: the news anchor's /rəpitən/³⁶ is likely a gerund . . .The word /fili/ "feel" appears in the question /wʌts du fili/, and is likely an infinitive (its lack of a final -en calls into doubt the possibility that /rəpitən/ is an infinitive.

I think I pulled a boner with this one you guys!!! The only way /fili/ would be an infinitive here is if /wʌts/ is q contraction involving something analogous to do-support; the fact that it ends with /-s/ which also serves as a 2sg verb ending could lend credence to this, but since there's no other evidence of do-support I'd say this is unlikely, and /fili/ is probably a regular 2sg conjugation, apparently in a verb class that uses -i instead of -s~z.

[Emily]

cleaned this up and expanded it on anthologi.ca