Official RPG Codex Discord Server

  1. Welcome to rpgcodex.net, a site dedicated to discussing computer based role-playing games in a free and open fashion. We're less strict than other forums, but please refer to the rules.

    "This message is awaiting moderator approval": All new users must pass through our moderation queue before they will be able to post normally. Until your account has "passed" your posts will only be visible to yourself (and moderators) until they are approved. Give us a week to get around to approving / deleting / ignoring your mundane opinion on crap before hassling us about it. Once you have passed the moderation period (think of it as a test), you will be able to post normally, just like all the other retards.
    Dismiss Notice

Linguistic comparison of IE games with Nu-Games

Discussion in 'General RPG Discussion' started by Bester, Feb 27, 2021.

  1. Bester ⚰️☠️⚱️ Patron Vatnik

    Bester
    Joined:
    Sep 28, 2014
    Messages:
    7,837
    Location:
    USSR
    To begin the comparison, I've extracted the NPC lines from the following games: BG1, BG2, IWD1, IWD2, PST, POE1, NUMENERA and TYRANNY.
    I've discarding the PC lines, because it would be unfair to the comparison - the player is not the story teller, he's often brief and to the point.

    I've then lemmatized the resulting words, meaning I've extracted their lemmas, which is the form in which a word would appear in a dictionary. This is not a perfect process. The alternative was Porter or Snowball stemming, which are also imperfect.
    After lemmatization, a sentence "Harry was in a better shape" will be turned into "Harry be in a good shape".

    The goal is to compare unique words (disregarding their varying forms) with TOTAL words, hence the procedure.

    Since these games are of a different length, I've decided to cut the comparison at a certain length which many of these games reach. IWDs and BG1 turned out too short to participate. POE1 barely made it.

    The results are, unsurprisingly, that Tyranny has the poorest vocabulary of all games, separated by a significant amount. Note that the Y-axis is exponential. It's fair to compare results exponentially, because a 10% difference is exactly what it takes for the text to become livelier and richer. Intersperse every 10th word with a rarer one and you get a smarter text. Tyranny is lagging behind by OVER 10%. It falls behind after 150k total words and never catches up. The difference only grows.

    [​IMG]




    For another comparison, I've extracted the most used words that are over 4 characters long.

    [​IMG]

    As you can see, all games are very subtle about their vernacular. PST is subtly charming with the heavy usage of "cutter", conveying the exotic setting with a multitude of other words that never repeat themselves. IWDs and NUMA emphase their most important toponym. BG1, 2 and even POE1 (yes, even POE1 despite its fampyrs) never verbally assault you with their unconventional world of magic and monsters.

    None of this can be said about TYRANNY which jumps out of the bush and rapes you with the MOST USED WORD fatebinder, followed closely by KYROS, ARCHON, CHORUS, DISFAVORED, BEASTWOMEN and many more. The writers don't even attempt to convey ideas elegantly. They've created a glossary and clumsily handle the narration by dropping the same words on your head over and over again.

    It's always been the author's opinion that Tyranny is an affront to any sane person's sensibilities. It has been proven by the numbers now. You can't like Tyranny and claim to be a respectable gentleman. Only a scullion could enjoy something like that.
     
    Last edited: Feb 27, 2021
    • Informative x 35
    • Brofist x 17
    • Salute x 10
    • Interesting x 4
    • (autism) x 4
    • Prestigious x 3
    • Participation Award x 2
    • Funny x 2
    • Acknowledge this user's Agenda x 1
    • hopw roewur ne x 1
    • nice x 1
    • incline x 1
    • Thanks! x 1
    • what? x 1
    ^ Top  
  2. rusty_shackleford Arcane

    rusty_shackleford
    Joined:
    Jan 14, 2018
    Messages:
    27,300
    I'm too lazy to download tyranny and analyze it myself so I'm going to assume it's bullshit until someone else confirms it
     
    • Edgy x 3
    • Balanced x 2
    • Funny x 1
    • Agree x 1
    • meh x 1
    • Shit x 1
    ^ Top  
  3. Tigranes Prestigious Gentleman Arcane

    Tigranes
    Joined:
    Jan 8, 2009
    Messages:
    9,734
    But POE1 is a curiosity here. Given the common criticisms, we would imagine that POE1 and Tyranny are similar with each other, and dissimilar with IE-games. But apparently this is not the case?

    One possibility here is that POE1 does feature a high quantity of 'vernacular', but it comes in the form of 800 different words, rather than a single one like 'cutter' dominating. That would require a secondary analysis to confirm one way or another, to see what is the overall proportion of 'jargon'. Without that, we can't yet conclude that Tyranny is guilty of jargon-dumping while all other games (including POE1) do not.
     
    • Agree Agree x 3
    • Disagree Disagree x 1
    ^ Top  
  4. Dodo1610 Magister

    Dodo1610
    Joined:
    May 3, 2018
    Messages:
    1,678
    Location:
    Germany
    Tyranny had this hyperlink feature where you could mouse over specific terms and you would get a popup explanation for it. I assume that that's why they kept using some terms over an over again

    Though I am surprised holowborn or Engwithan isn't on POE list.
     
    • Agree Agree x 8
    • :M :M x 1
    ^ Top  
  5. laclongquan Arcane

    laclongquan
    Joined:
    Jan 10, 2007
    Messages:
    1,868,838
    Location:
    Searching for my kidnapped sister
    INteresting.

    1.Note: Bloom word in Numa is a specific usage there, or just normal meaning?
    2. Also voices, and disfavoured in TYR. is that normal word, or is there specific uncommon meaning to them in Tyr?
     
    • Yes Yes x 1
    ^ Top  
  6. Bester ⚰️☠️⚱️ Patron Vatnik

    Bester
    Joined:
    Sep 28, 2014
    Messages:
    7,837
    Location:
    USSR
    Hollowborn top 336 most used word with just 83 usages.
    Engwithan top 351 with 80 usages.

    They didn't overdo it. I couldn't even remember those words before you mentioned them.

    "Voices of Nerat" is some dude and Disfavored is a faction.

    A place.
     
    • Informative Informative x 1
    ^ Top  
  7. Bester ⚰️☠️⚱️ Patron Vatnik

    Bester
    Joined:
    Sep 28, 2014
    Messages:
    7,837
    Location:
    USSR
    You are right, but I'm not doing it.

    While it's possible to discard all words that exist in dictionaries, we'll be left with a lot of stuff like this from Torment for example (the undead language), and then it'll be a matter of manually going through all of this. It's not a particular case, there's a lot of stuff like this in these games, so I'd rather not.
    [​IMG]
     
    • Funny Funny x 3
    ^ Top  
  8. Sweeper Arcane Zionist Agent

    Sweeper
    Joined:
    Jul 28, 2018
    Messages:
    2,292
    How do I get the autism button?
     
    • Equip the Ring x 4
    • (autism) x 4
    • it is a mystery x 3
    • Participation Award x 2
    • Edgy x 1
    • Old x 1
    • /facepalm/ x 1
    ^ Top  
  9. Pegultagol Erudite

    Pegultagol
    Joined:
    Feb 4, 2005
    Messages:
    1,164
    Location:
    General Gaming
    I would be interested to know if there's any appreciable difference between POE1 and POE2. And if the number of writers involved in the project affect the scope of common lemmas in the dialogue.

    It was nonetheless a very interesting read. Thanks Bester.
     
    ^ Top  
  10. AdolfSatan Magister

    AdolfSatan
    Joined:
    Dec 27, 2017
    Messages:
    1,226
    It shouldn't as long as there's a good lead writer in charge of the team. So yes, it probably does.

    Interesting idea for a thread, but it feels a bit half-baked. Neither lack nor excess in verbosity implies good quality on itself, it'd be interesting to see it followed by some other criteria of analysis. Otoh it's kind of pointless since a quick glance is enough to realize which games have shit writing.

    Any chance you could add other games to the list? It'd be interesting to see what the graph looks like for BaK, for example.
     
    • Agree Agree x 1
    ^ Top  
  11. ERYFKRAD Barbarian Patron

    ERYFKRAD
    Joined:
    Sep 25, 2012
    Messages:
    19,395
    Serpent in the Staglands Shadorwun: Hong Kong Pillars of Eternity 2: Deadfire Pathfinder: Kingmaker
    Fascinating.
     
    • Funny Funny x 3
    • Agree Agree x 2
    • No No x 1
    • :M :M x 1
    ^ Top  
  12. Bester ⚰️☠️⚱️ Patron Vatnik

    Bester
    Joined:
    Sep 28, 2014
    Messages:
    7,837
    Location:
    USSR
    Propose a criterion. If I think it makes sense, I may run it.

    I was thinking of how the Chinese and the Japanese have categorized their words by age at which these words must be learned at school, because each year they need to learn specific new hieroglyphs. So if I was analyzing a Japanese test, I'd be able to tell the age of the audience the text was aimed at. But I don't think there are corpora (linguistic resources, such as dictionaries) that have assigned a rarity value to words in English. At least I couldn't find it after a few minutes of googling.

    It's possible to find the average word length just for fun, but it's probably not a serious criterion for anything. Do you have anything in mind?
     
    Last edited: Feb 27, 2021
    ^ Top  
  13. sser Arcane Cuck Developer

    sser
    Joined:
    Mar 10, 2011
    Messages:
    1,866,466
    Unironically cool.

    Also found an amusing horizontal synchronicity here:

    upload_2021-2-26_23-41-57.png
     
    • Funny Funny x 2
    • Yes Yes x 1
    • it is a mystery it is a mystery x 1
    • Balanced Balanced x 1
    ^ Top  
  14. ERYFKRAD Barbarian Patron

    ERYFKRAD
    Joined:
    Sep 25, 2012
    Messages:
    19,395
    Serpent in the Staglands Shadorwun: Hong Kong Pillars of Eternity 2: Deadfire Pathfinder: Kingmaker
    Think you can add New Vegas and AoD to the mix?
     
    • incline incline x 2
    ^ Top  
  15. Zibniyat Arcane

    Zibniyat
    Joined:
    Jun 22, 2014
    Messages:
    4,890
    Fatebinder, Archon(s) Chorus, Kyros, edict(s), Tiers etc. are all unique terms for either certain persons (for example Voices of Nerat, hence voices), locations (Tiers), or factions (the Disfavoured).

    Yes, they do repeat often, but the game presents a conflict between factions, with the player being simply part of it, hence the continued use of these same unique terms.
     
    ^ Top  
  16. Tigranes Prestigious Gentleman Arcane

    Tigranes
    Joined:
    Jan 8, 2009
    Messages:
    9,734
    Yeah, I understand why you wouldn't want to bother. It just means that we can't really answer that particular question of whether Tyranny is more jargon-abusive than POE1 or IE games. What we do know is that Tyranny is very focused in its jargoning - which is interesting!

    (Another side effect of this issue: it's surprising that Sarevok is up there but not Bhaalspawn in either game. And maybe that's because it gets split into "Child of Bhaal", etc.)

    I don't know what the order of the data was when you cut it, but we don't know if that affected the results either, e.g. a game being more/less jargony at the start as opposed to later.
     
    • retadred retadred x 1
    ^ Top  
  17. laclongquan Arcane

    laclongquan
    Joined:
    Jan 10, 2007
    Messages:
    1,868,838
    Location:
    Searching for my kidnapped sister
    The fact TYR use a bunch of unique terms certainly can be used as proof of "jargon-abusive".
     
    ^ Top  
  18. KK1001 Arbiter

    KK1001
    Joined:
    Mar 30, 2015
    Messages:
    620
    Do Disco Elysium next.
     
    • Agree Agree x 1
    ^ Top  
  19. AdolfSatan Magister

    AdolfSatan
    Joined:
    Dec 27, 2017
    Messages:
    1,226
    I hadn't thought of anything in concrete, but I'll throw some ideas. No idea how feasible they are.

    How about an algorithm that pulls out entire sentences and categorizes the words within, replacing them with their type (noun, pronoun, verb, adverb, adjective, etc.). You can either sum and dump that info, or process it again to get data on how the average sentence is constructed. Once more, cold data isn't everything since good taste may be found both on sparse and overflowing prose, but it'd be an interesting take.

    On that same page, you could make data dumps for lemmas on specific categories as well.

    Another thing you can do is count the amount of adjectives and pool it against lazy words (very, little, rather, quite, pretty, a lot, more, etc.). This should account on how rich/poor in adjectives the vocabulary is.

    Flow and pacing is important too, how about pulling paragraphs and getting info on how many words and sentences form them? You can then average the sentence/paragraph length, or expand it further by setting length categories for sentences (say, 3 levels) and analyzing their distribution inside paragraphs. I have no idea how to average that last one though.

    And one that's lazier to program but demands crossing data online might be drawing a set amount of random samples from each game, sending it to grammarly (or any other site) and getting a count for how many errors it returns and on which categories.
     
    ^ Top  
  20. Beowulf Arcane

    Beowulf
    Joined:
    Mar 2, 2015
    Messages:
    1,558
    It's the same game, but 2x biggerr:
    [​IMG]
     
    • Funny Funny x 13
    • :M :M x 2
    ^ Top  
  21. CappenVarra phase-based phantasmist Patron

    CappenVarra
    Joined:
    Mar 14, 2011
    Messages:
    2,812
    Location:
    Ardamai
    :hmmm:

    excluding BG1 and IWD1 is a mistake, cut things off at a smaller threshold to include them or use relative counts instead of absolute ones or...
     
    ^ Top  
  22. Funposter Liturgist

    Funposter
    Joined:
    Oct 19, 2018
    Messages:
    1,220
    Location:
    Australia
    There's stuff like this floating around. Phrases such as "reading at an 8th grade level" etc. are relatively common in American television and writing when dealing with the subject of education, although I don't know if I've ever heard of attempts to do something similar in other English speaking countries such as the UK or Australia. There was also some talk about the educational level that US Presidents and presidential candidates spoke at back in 2015/6 due to Trump reportedly speaking at a "fourth grade level". In this example, it should however be noted that the educational level of Presidential speeches has been trending downward since the mid-19th century.
     
    ^ Top  
  23. rusty_shackleford Arcane

    rusty_shackleford
    Joined:
    Jan 14, 2018
    Messages:
    27,300
    only thing I learned from writing papers in grad school is that communication should be easy, not hard
    midwits who like to huff their own farts disagree with this though
     
    • Agree Agree x 5
    • Disagree Disagree x 2
    ^ Top  
  24. Funposter Liturgist

    Funposter
    Joined:
    Oct 19, 2018
    Messages:
    1,220
    Location:
    Australia
    Agreed but there's probably a point at which simplifying speech for ease of communication turns into catering to retards/blacks.
     
    • Racist Racist x 2
    • hopw roewur ne hopw roewur ne x 1
    • Shit Shit x 1
    ^ Top  
  25. Rahdulan Arcane Patron

    Rahdulan
    Joined:
    Oct 26, 2012
    Messages:
    4,437
    To be fair, it would be interesting to see how many of those words associated with Tyranny come from hyperlink text rather than from dialog proper. Those can be totally ignored by the player.

    Damn, Dodo1610 beat me to it.
     
    ^ Top  

As an Amazon Associate, rpgcodex.net earns from qualifying purchases.