Talk:2942: Fluid Speech

2024-06-07T17:11:45Z

172.69.70.139: Proposed sample video for nazalized schwa

I've noticed that this doesn't seem to be the case in all languages. For example, when native Polish speakers talk rapidly (even when speaking English), they enunciate every sound accurately in quick succession while flattening out the tone and rhythm of their speech. I wonder if this is because Polish is an inflected language where the grammar of the sentence is determined by endings of words rather than word order. Does anyone know if there have been any studies on this? [[Special:Contributions/162.158.74.49|162.158.74.49]] 23:12, 5 June 2024 (UTC)

:I'm not linguists but based on how many those are, definitely. -- [[User:Hkmaly|Hkmaly]] ([[User talk:Hkmaly|talk]]) 00:10, 6 June 2024 (UTC)

:Russian also has vowel reduction like English and it's a Slavic language like Polish, so I don't think so. Although someone who knows more than me might be able to chip in on whether the effect is stronger in English. [[Special:Contributions/162.158.114.198|162.158.114.198]] 03:24, 6 June 2024 (UTC)

::Russian vowels are a tangle of reduction indeed. I think they’re much more complicated than English vowels, but I may be biased as a native speaker of Russian who started self-studying phonetics, phonology etc. to improve his English pronunciation and only relatively briefly looked up Russian phonology for fun. But at any rate, Russian vowel reduction happens in slow and fast, formal and colloquial speech alike, pretty much universally except when articulating a word exceedingly clearly when someone can’t hear you well; and Russian consonants are generally unaffected outside of several specific clusters and morphemes, even if you include those that are fully codified in modern language but retain etymological spelling (чувство, счастье, солнце). [[User:Chortos-2|Chortos-2]] ([[User talk:Chortos-2|talk]]) 16:30, 6 June 2024 (UTC)

:I know the comic specifies native speakers, but I just asked some of my East Asian friends and they very clearly enunciated the "t" in "hot potato". [[User:Tcf|Tcf]] ([[User talk:Tcf|talk]]) 07:22, 7 June 2024 (UTC)

I've just added a very barebones version of an explanation based on what I could understand from the comic. I can tell that the four diagrams depict that of the human mouth but since I am not a linguist, I lack the knowledge of various terms and thus, can't fully explain the comic. I understand what the comic is trying to convey, I just can't explain it. Looking forward to seeing how this progresses. [[User:OmniDoom|OmniDoom]] ([[User talk:OmniDoom|talk]]) 00:22, 6 June 2024 (UTC)

I don't think #4 is a real IPA symbol, but as I am not a linguist, I have no idea. [[Special:Contributions/162.158.91.36|162.158.91.36]] 01:38, 6 June 2024 (UTC)
:/ə̃/ is a {{w|Nasal_vowel|nasalized}} {{w|schwa}} --[[Special:Contributions/172.71.160.92|172.71.160.92]] 08:53, 6 June 2024 (UTC)
:A video for demonstrating the sound: https://www.youtube.com/watch?v=0Q2M9ILulTo [[Special:Contributions/172.69.70.139|172.69.70.139]] 17:11, 7 June 2024 (UTC)

The hovertext joke is that every English speaker fully pronounces the first "t" in "Hot Potato". It's at the end of "hot". Nobody says "ha potato". [[User:Nitpicking|Nitpicking]] ([[User talk:Nitpicking|talk]]) 03:01, 6 June 2024 (UTC)

:Unless you mean "the glottal stop should be considered an allophone of <t> at the end of syllables" then yes they do. It's /hoʔ/, not /hotʰ/. [[Special:Contributions/162.158.114.198|162.158.114.198]] 03:24, 6 June 2024 (UTC)

:(Edit conflicted with 162, above, so this is my own reply...) I thought it was that it isn't "ho T'p otato", with the "teh-peh" awkwardness. For me, the natural way to say it is to glottalstop the first T for more "ho'potato" (the other Ts, there I find awkward ''not'' to get the "t>s<" out of, the ">s<"-tail being what makes a full-T not a lazy one). But clearly a different accent involved, as "ha" doesn't work at all for me unless I try to use some sort of (probably awful) Goodfellas-type accent. And my native accent is notoriously good at glottlestopped Ts (that most people misinpersonate badly, by attaching them to the wrong adjacent syllable).
:As for "going to", experimentally holding my finger over the length of my tongue, it seems it barely has to move at all in "going" (the whole tongue wants to rise on the "i", but I can suppress that and do the tone-change from further back, if not straight from the vocal chords). Though continuing through to the "to", with my finger in there, it's no better than "going ku" as I prevent the tongue-tip doing the necessary small movement to fulfil any form of T. I can do better through basic gastromancy, but behind my unmoving jaw and lips (''without'' the finger almost down my throat, of course), I can feel the tongue tip doing it's small but vital "crossing the T" work. [[Special:Contributions/172.70.86.64|172.70.86.64]] 03:53, 6 June 2024 (UTC)
:For me, it's more like "hobuh-deh-duh" - so none of the t's get pronounced properly. And I'd drop the n in 'going to' before I dropped the t.[[Special:Contributions/172.70.163.120|172.70.163.120]] 08:22, 6 June 2024 (UTC)
:Wunber dayder, doober dayder, freeber dayder for. (But then, oddly, uh baguv tayterz.)[[Special:Contributions/172.70.162.58|172.70.162.58]] 13:18, 6 June 2024 (UTC)

Unless someone's willing to start an "explain explain xkcd", I think this explanation still needs a lot of work to be intelligible to non-linguists (myself included). That aside, I do appreciate whoever took the time to type all that up. [[Special:Contributions/162.158.166.210|162.158.166.210]] 03:31, 6 June 2024 (UTC)
:The current explain reads, to me at least, more like a 102 lecture than an explanation of the comic. I of course have no idea what is in a 101 first week lecture so shrug. (Aside, wth? This keyboard doesn't have a tilda. Copy and paste ftw) [[Special:Contributions/172.71.223.38|172.71.223.38]] 05:51, 6 June 2024 (UTC)
:::Meta-response about the {{w|tilde}}: at the top of this editing window is a bar of formatting buttons (which I mostly just try to avoid accidentally clicking when I touchscreen-scroll). The second from the right gives <code><nowiki>--~~~~</nowiki></code>, having a depiction of a signature. (The "--" is not necessary, nor does anything with actual formatting/markup, but comes from quite old text communications standards.) Personally, in this current situation of using an onscreen keyboard, my configurstion hides the tilde behind the "?123" then "=\<" change-keyboard buttons.
:::(Ironically, all three of "=", "\" and "<" are already available as long-presses of the primary keyboard layout. But the much more useful "/" is hidden behind the "?123" press, ''except'' when it explicitly detects that I'm in a browser address field. Some UI designers have strange ideas that definitely mis-mesh with my usage!)
:::That's where I usually go, to sign-off. But on physical keyboards, depending upon internationalisation options, it might be either off the top-left (left of "1" key) 'triple-key, perhaps needing Shift or AltGr fingering (from experience of US (mis)configuring), or the key in the "hook" of the <Enter> key (all my physical UK keyboards, even the most squished-up laptop ones, have that as <sub>#</sub><sup>~</sup>). If you're neither in US nor UK (and your device knows this), then where it gets shuffled out of the way of any ß, ē or ø type stuff, I wouldn't know for sure, but using the AltGr (right-Alt) ''might'' reveal characters you never ever knew you had... ;)
:::Copypasting out of the "<nowiki>Please sign your posts with ~~~~</nowiki>" Infobox or any residual from the {{template|unsigned}} templates is, of course, also a valid option. ;) [[Special:Contributions/172.69.43.184|172.69.43.184]] 10:18, 6 June 2024 (UTC)
: I think it's fine if you can read/interpret the {{w|International Phonetic Alphabet|IPA}}. If you don't it's utterly incomprehensible. I think we need some examples here as to how the sounds written here are pronounced. Like "sound <x> as in <word>" [[User:Elektrizikekswerk|Elektrizikekswerk]] ([[User talk:Elektrizikekswerk|talk]]) 06:52, 6 June 2024 (UTC)
:: Yeah, except [[2819]] [[Special:Contributions/172.70.163.120|172.70.163.120]] 08:30, 6 June 2024 (UTC)

>Alien impersonating a human

Sounds like a normal Runglish to me, just like the one you can hear in this clip: [https://www.youtube.com/watch?v=GXNKUo5MrbM]

I hate that he pointed this out, because I'm going to start pronouncing things the way they're supposed to sound and everyone will think I'm weirder. [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 12:37, 6 June 2024 (UTC)
: How are you pronouncing 'Psychoticpotato'?[[Special:Contributions/172.69.195.6|172.69.195.6]] 13:10, 6 June 2024 (UTC)
::Think of the words "psychotic" and "potato". [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 15:35, 6 June 2024 (UTC)

I added a bit more detail to the first diagram, but I agree it's not really plain English right now. btw, maybe it's pedantic of me, but diphthongs such as /oʊ/ are one vowel, and the whole word /ɡoʊ.ɪŋ/ only has two vowels

As for "hot potato", the post-vocalic T *is* pronounced, it's pronounced as a glottal stop. T is simply pronounced differently in different positions, it's how the letter works. Randall's probably referring to the alveolar plosive /t/, which most people think of as "the T sound" and would make you sound like an alien. SMBC made the point better: https://www.smbc-comics.com/comic/2012-05-08
[[User:FrustratedPhonetics|FrustratedPhonetics]] ([[User talk:FrustratedPhonetics|talk]]) 13:57, 6 June 2024 (UTC)
:Unless, like me, you just elide straight into the following consonant. Hock ross buns, hock ross buns, wunner penee tooer penee, hock ross buns.[[Special:Contributions/172.69.43.166|172.69.43.166]] 14:19, 6 June 2024 (UTC)
::This causes me physical discomfort. [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 15:41, 6 June 2024 (UTC)
:I find I say 'hod p'tato' but the 'd' only the precursor to 'd' that's never 'plosived': my tongue forming a tube with the roof of my mouth, closed just behind my teeth, so I don't think that's glottal. [[Special:Contributions/172.70.85.168|172.70.85.168]] 10:21, 7 June 2024 (UTC)

Not exactly original, "hopotado": that's just how French works. Consonants at the end of a word are only pronounced if the following word starts on a vowel. In which case the two are drawn into one word. The cool thing about French is that knowing this little rule, you can read out a text alound and others understand even if you don't. :D --[[Special:Contributions/162.158.86.100|162.158.86.100]] 21:38, 6 June 2024 (UTC)

A very small proportion of people can read and understand the IPA, and the explanation depends on knowledge of it currently. So it's a terrible explanation which needs to be written without phonetician jargon. [[Special:Contributions/162.158.42.81|162.158.42.81]] 05:16, 7 June 2024 (UTC)

The UK "bo'le" - I find I still say the 't' but with the sides of my tongue instead of the tip. [[Special:Contributions/172.70.85.168|172.70.85.168]] 10:21, 7 June 2024 (UTC)

If you don't fully touch your tongue to the roof of your mouth when you say "hot", it sounds normal while also having the "t" sound. I learned this by saying the phrase multiple times in class and having people look at me strangely. [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 12:36, 7 June 2024 (UTC)

== "Sandhi" ==

Now I want someone to name their kid Sandhi, pronounced [säní].
[[User:ProphetZarquon|ProphetZarquon]] ([[User talk:ProphetZarquon|talk]]) 14:17, 7 June 2024 (UTC)
:Poor kid. Every substitute teacher they had would pronounce it wrong. [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 15:42, 7 June 2024 (UTC)

== Found while reading ''Buttercup Festival'' ==
Grade A gray day!
[[File:grade_a.jpg|frameless|caption]]
-- [[User:Psychoticpotato|Psychoticpotato]] ([[User talk:Psychoticpotato|talk]]) 16:07, 7 June 2024 (UTC)

2731: K-Means Clustering

2023-01-30T15:42:01Z

172.69.70.139: /* Explanation */

{{comic
| number = 2731
| date = January 30, 2023
| title = K-Means Clustering
| image = k_means_clustering_2x.png
| imagesize = 320x385px
| noexpand = true
| titletext = According to my especially unsupervised K-means clustering algorithm, there are currently about 8 billion types of people in the world.
}}

==Explanation==
{{incomplete|Created by EITHER 8 BILLION OR 3 TYPES OF BOTS - Please change this comment when editing this page. Do NOT delete this tag too soon.}}

[https://en.wikipedia.org/wiki/K-means_clustering ''k''-means clustering] is a method of categorising ''n'' vectors into ''k'' clusters. For example, we might categorise a population by two metrics, and want to best categorise this scatter graph into the distinct populations, algorithmically drawing [https://en.wikipedia.org/wiki/Voronoi_cell Voronoi cells] to decide the within-cluster variances.

The title text refers to a K-means algorithm applied to the human population, which has, as yet, not converged to assign any two human beings into a common cluster based on shared traits. This is humorous because it is effectively saying that each individual is an entirely unique individual, which would make such a clustering useless as there is literally no commonality between any two people so you can't make any cohort out of any two people for the purposes for which a K-Means Clustering is typically used, such as of making insurance risk pools or targets of advertisement campaigns.

Interestingly, by including the entire human population, the algorithm should be immune to bias in creating its input data. However, since every human is unique, the only way to have the clusters converge is to "throw out" some traits of humans as unimportant. This may be objectionable to humans who disagree with that assessment. In contrast, in a supervised algorithm, the training data is tagged with traits that the trainers seek. These traits could be applied in a manner that is socially unacceptable, and lead to AI behavior that reflects the biases of the trainers.

==Transcript==
{{incomplete transcript|Do NOT delete this tag too soon.}}
:Ponytail is presenting on a stage, pointing a screen with a stick. The writings and possible figures on the screen are illegible.

:Ponytail: Our analysis shows that there are three kinds of people in the world: Those who use ''k''-means clustering with ''k''=3, and two other types whose qualitative interpretation is unclear.

{{comic discussion}}
[[Category:Comics featuring Ponytail]]

explain xkcd - User contributions [en]

Talk:2942: Fluid Speech

2731: K-Means Clustering