Who Will Speak?

It's never entirely spelled out in classic D&D exactly which creatures have their own racial language. OD&D says (Vol-1, p. 12), "The 'common tongue' spoken throughout the 'continent' is known by most humans. All other creatures and monsters which can speak have their own language..." But which ones are those?

When my players have the opportunity to learn a new language, I'm in the habit of handing them the list of aligned creatures from that same book (Vol-1, p. 9), but the more I've used it, the more some of those types look suspect to me:

Now, the AD&D game has more clues. The DMG (p. 102) has a random list of languages, so we can firmly settle for that game edition that all these creatures speak their own language:

Note the last category (86-00) and its associated footnote; "other" can be any "unlisted creature language", so apparently appearance in this table is only sufficient, but not necessary, for a race to have its own language.

We know that Intelligence is related to one's ability to speak a language (that's the ability that sets the number of languages one can know in every edition). Let's say we scan though the AD&D Monster Manual and see if we can correlate creature Intelligence to speaking a language.

We find this: There are a half-dozen monsters who have their own language (per the DMG table above), and who have an Intelligence in the category of "Low" (meaning 5-7 points per the MM preface). These are: Ettins, Giants (Hill), Manticores, Minotaurs, Ogres, and Trolls. No creature in that table has an Intelligence any lower than that, with one exception; the Chimera is listed as Semi-Intelligent (2-4 points), and of that monstrosity it is written, "Chimerae speak a very limited form of red dragon language" (MM p. 14).

So we might draw the pattern that creatures of Low Intelligence are most likely to speak their own language, whereas creatures with Intelligence below that usually don't speak, or at best, use a primitive pidgin version of some other race's. Based on that, it seems a fair bet that these other races also have their own languages in either edition, based on listed Intelligence in the MM: Treants (Very intelligent), Unicorns (Average), Pegasi (Average), and Wyverns (Low).

Now, here are some types that would seem to have lower Intelligence in AD&D than would permit language use: Rocs (Animal), Hippogriffs (Semi-), and Gorgons (Animal). However, note that AD&D has definitely made a big change for each of those in at least one other aspect; whereas in OD&D they are all aligned (Rocs* and Hippogriffs being Lawful; Gorgons being Chaotic), in AD&D they are instead listed as Neutral. To me, this seems to mark a change in those creatures who were formerly seen as being sentient (some minimal intelligence required to be non-neutral?). So back in OD&D I would read their being aligned as implying intelligence, and therefore also language-speaking. Likewise (since Pegasi and Hippogriffs both presumably speak), I would take the Griffon as also being also able to do so, even though it is Neutral in both editions.

To some extent OD&D echoes in this way the fantastic freedom in The Hobbit that almost any creature seems to possess intelligence and some unique language, whereas AD&D and Lord of The Rings seemed to get more restrictive/mundane in this regard.

This leaves a very small number of monsters from the OD&D list as possibly not speaking any language, creatures that are Neutral in every edition, and also have very low Intelligence in the Monster Manual: Hydras (Semi-Intelligent), Purple Worms (Non-Intelligent), and Sea Serpents (as a Plesiosaurus per Arneson in Sup-II; Non-Intelligent in the MM). Now, Hydras have the same Intelligence as Chimeras, so I'm prone to say they can speak, perhaps again in a degraded form of Dragon or whatever. The entirely Non-Intelligent types (which means 0 points per the MM), the Purple Worm and Sea Serpents, I'm pretty comfortable as leaving non-speaking mindless monsters. But that's (surprisingly) only two out of the entire OD&D list!

I guess (having mentioned The Hobbit), I should address the "Animals" entry in that same OD&D list. Certainly I wouldn't want a single language for all "Animals" (as some new players of mine have hoped for, given the way I introduced that table). Per The Hobbit, maybe every animal type has its own language. But following the Monster Manual, I think I'd require the type to have above "Animal Intelligence"; I'd be comfortable with any giant type so speaking?

* Finally: Note the particularly interesting evolution of the Roc in D&D. In the earliest printing of Chainmail, it is expressly written that, "These equal the 'Eagles' of Tolkien's Trilogy". And again in the first printing of OD&D (Vol-2): "This term has been used to encompass large and fierce birds such as the 'Eagles' of Tolkein", as well as expansion on their strong sense of alignment: "they will be positively hostile only to Chaos and Neutrality, ignoring (80%) or being friendly (20%) to Lawful characters".

In contradistinction, AD&D seems to split these types apart, with a separate entry for Giant Eagles (MM p. 36) with Average Intelligence and who "have their own language", versus Rocs (MM p. 82) who are now switched to Neutral and given a lowly Animal intelligence (and hence presumably non-speaking in that edition, by my analysis). In OD&D I'm happy to keep the original intent as sentient Tolkienian Eagles intact (and also not multiply entities needlessly).


  1. It bugs the hell out of me that every subspecies has its own language; that every one of the nine-point alignments has its own language; that goblins somehow speak a different language from hobgoblins.

    I'm of a mind to make up a list of most-common languages in order to allow my players to have even a tiny chance of picking something broadly applicable.

    What do you think? Maybe 15 languages in total? I don't know. Something in that range.

    1. Here's the list I just came up with. Can anyone suggest more?

      Race Languages:

      1. Dwarven (gnomish is a dialect.)
      2. Elven
      3. Goblin (goblins, hobgoblins and bugbears)
      4. Draconic (dragons and similar creatures with linguistic capacity)
      5. Giant (including ogres and trolls)

      Alignment Languages:

      6. Lawful
      7. Achromatic (Neutral Alignment)
      8. Chaotic

      Outsider Languages:

      9. Celestial (Good)
      10. Infernal (Lawful)
      11. Abyssal (Chaotic)

      Infernal is very, very, very specific. It is the language chiefly of Devils. It is rarely if ever spoken by Men of the Realm but important contracts are written in Infernal as well as Common and/or Dwarven.

      Celestial and Abyssal are superglottal languages. Men have not the ability to speak it, though they may learn to understand their aural form. Audible Celestial forces evil creatures to check morale or cower in awe. Audible Abyssal does the same for Lawful creatures. (Note: A creature may be both Lawful and evil.)


      12. Common (men and hobbits)
      13. Sylvan (forest creatures)
      14. Undercommon (kobolds and subterranean beings)

      Secret Languages:

      15. Thieves' Cant
      16. Druidic

      Secret Languages are not available to most characters at creation except for those characters who are part of the organizations which use them.

    2. There's a strong case for that; it's very similar to what published D&D did, starting with the 3E. Link.

      That said, my own game is chugging along pretty well with this wide array of languages on the table. The OD&D rule is really liberal with languages: 1 extra per point of Int over 10 (AD&D being similar). Looking at my top 5 PCs right now, between them they know 19 exotic extra languages. It's kind of nice that different players get put on the spot to try and parley with monsters in different encounters.

      That said, I am about to finally kick the alignment languages to the curb. Those are just too problematic in play, and having recently read that at Gygax's table the whole issue was short-circuited by allowing PCs to learn all 3 alignment languages, that put a real foul taste in my mouth for them.

  2. If you check out the 3rd edition, they pretty much did this canonically. All giants speak Giant. All dragons (and I think Kobolds) speak Draconic. All earth-element or burrowing creatures speak Terran (Or Ignan for fire-, Aquan for water-, and Auran for air-elementals). Animals that one talks to by magic usually speak Sylvan. Demons and Devils both speak Infernal (though the Devils insist the Demons are doing it wrong), and celestials speak Celestial.

    At the same time, Sylvan is basically forest- or Elf-common, Giant is giant-common, etc.

    It really pares down the list and makes it manageable... but if languages are manageable that really diminishes their value. A wizard with good intelligence can at first level reasonably say that he's studied all the obscure languages as part of his magical studies, and the more common languages are filled in by the rest of the group. They never really matter. It's sort of an illusion of support.

    1. Is it your opinion languages rather ought to be more granular instead of more universal? Is that right?

    2. My opinion is that if you have an element that looks like it should matter, it should matter. Otherwise, don't have it. Languages look like they should matter, but rarely do in D&D 3+e. Further, if you want to try to make languages matter in your campaign setting, the 3e approach coagulates them so much that they really won't matter anyway.

      Now, a somewhat heavy-handed e DM could rule that certain languages are unlearnable without some adventuring, such as Infernal or Ignan. But those languages are least likely to be worth anything anyway, because those creatures either don't want to negotiate or can speak common.

  3. The OD&D Roc/Eagle was also inspired by the Tarns of Gor, which Gygax never seems to have cared for, but Arneson certainly did. Splitting them into aligned Giant Eagles (per Tolkien) and unaligned beasts (per Norman) was probably to disentangle them a bit.

    1. Okay. I see in Arneson's First Fantasy Campaign that Rocs and Tarns are separate types in his transportation price list.

  4. I kind of fall back on the a comment I made about alignments a while back.
    Lawful creatures "make" things, so lawful/civilized creatures "make" languages too. Monstrous beings probably speak a pidgin version of some other language (or common which i treat as a pidgin language and not a true language)
    Plus, the PCs having some chance at communication (with the possibility for misunderstandings) i think opens up some gameplay options.
    Frodo and company had no issue speaking to Ogres, Goblins etc. Ravens too (Though I always imagined the ravens speaking dwarven....)

    1. That's an interesting take on it. OD&D gives a 20% chance that any speaker knows Common, so in my head I interpret Tolkien's scenes and having found someone in that category. Meanwhile, knowing their language helps to eavesdrop, read runes, etc. But your idea is neat.

  5. I’m glad you pointed out the Tolkien Roc/Eagle thing so that I wouldn’t have to. Clearly the giant roc of Arabian folklore has a place in fantasy adventure games, but it is hugely distinct from the JRRT Eagle, always being portrayed as a dangerous, giant beast of (basically) bird intelligence and natural (animal) instinct rather than something sentient. I imagine that Norman used the roc as inspiration for his tarns, as there is a lot of “oriental exoticism” in his books (including an assassin caste, which also makes its way into Blackmoor and hence D&D).

    Ha! Sometimes parsing our the origins of this game is akin to parsing out the evolution and migration of language families, huh?
    ; )

    1. Absolutely!

      I'll also add for those who don't know that OD&D starts rocs at 6 HD and then the text specially says they could be double or triple sized (which would match the AD&D MM), so maybe there's an implied distinction between the small and larger types there.

  6. I diverge from your way of handling languages in two seemingly opposite ways.

    For civilized folk, I want more diversity in languages, without necessarily adding literal languages. My simple way of handling this is to link language differences to distance along a river, or overland to the closes river settlement. The farther you go up or down stream, the more problem you will have communicating with the locals. Originally, I had modifiers based directly on distance, but I'll probably simplify that.

    For non-human sentients, I actually take away some languages and make them speak corrupt dialects of civilized languages. The goblinoids all speak corrupt dwarvish, which can be understood with difficulty by those who speak dwarvish. But at the same time, I kind of like a mild fairy tale quality, so I'd allow animal languages... but rather than each species getting its own language, I'd probably break it down by biome, and humans can't "speak" it, only understand it. And I probably wouldn't let a PC start with those languages without good reason.

    The big picture result is that there's a lot more difficulty in communicating. It's not too bad, though, because I also treat the languages based on Int as starting languages known, not max languages known. I leave that open-ended, but speed of learning languages is affected by Int.

    1. Interesting, the "corrupt dialects" part sounds a bit like Baquies' approach above.

  7. Let me see...
    d20-based games usually have a Speak Language skill (D&D 3e/3.5, Star Wars RPG, Pathfinder), but some also had a separate Read/Write (Other) Language (d20 Modern, Call of Cthuhlu d20). In d20 Modern, they weren't languages so much as language groups (like the Celtic, Germaninc, Iranian, or Semitic languages). Most of these use a "you know it or you don't" system, though d20 Modern suggested you could make a check to see if your general understanding of e.g. Semitic languages was sufficient to understand a text in Akkadian specifically. In most of the d20 games, you could also treat written languages a code you could decipher with Decipher Script skill. These games also called out which language was in which written script (with Elvish and Sylvan sharing a common script, Dwarvish and Terran another, and so on), but did not offer much in the way of why is should matter.

    AD&D 2e had some language compatibility rules. The Greyhawk boxed set had some -rules- percentages for mutual intelligibility. For example, that the Keolandish language is 60% compatible with both the Oeridian and Common. Pretty much everything was a 60%, 40%, or implicit 0% compatibility. There's no clear rules for how you're supposed to apply those percentages.

    AD&D 2e Forgotten Realms sets played with some fringe language compatibility rules. For example, Drow of the Underdark mentioned that Deep Drow (i.e. Low Drow) language is a descendant of High Drow (i.e. Elvish) with loan-words from Undercommon (i.e. a dialect of surface common, which is to say, Thorass) such that one who knows BOTH has ~14% chance of understanding it. Giantcraft has a table of language proximity for "Jotun" (Giant Common) and its descendant languages (Jotunalder, Jotunhaug, Jotunise, Jotunild, Jotunskye, Jotunuvar).

    The most complicated thing I've seen was in the Forgotten Realms: The Horde box set, where the various languages of the Steppe were laid out in a map: each pip on the path connecting two languages represented a further -10% to language overlap. So for example, a six-pip separation mean the two step languages had a 40% overlap. Again, very vague on how this was to be applied.

    1. Good gravy! But then the Horde approach looks a lot like Talysman's above.

      For simplicity, I kind of don't mind OD&D's binary yes-or-no take on knowing a language. But the d20 cost of Speak Language skill (1 point for full fluency) looked way too easy in the context of that system.

  8. Instead of the alignment chart, maybe better to just go through the monster list on vol 2 pp. 4-5 and number the ones you want. I'm not as inclined toward animal languages, so I think I'd end up with something like

    1. Goblin
    2. Kobold
    3. Orc
    4. Hobgoblin
    5. Gnoll
    6. Ogre
    7. Troll
    8. Giant
    9. Manticore
    10. Dragon
    11. Demon (i.e. "Balrogs")
    12. Minotaur
    13. Centaur
    14. Nixie
    15. Pixie
    16. Dryad
    17. Gnome
    18. Dwarf
    19. Elf
    20. Roc

    With a couple of Greyhawk replacements—Bugbear instead of Nixie, Lammasu instead of Manticore, Lizard Man instead of Dryad or Roc —it could make for a nice simple list of 20.

    1. That's not bad. FWIW, in my copy of Vol-1 p. 9 I've got the separate columns independently numbered for dice-rolling (the Chaos column alone has 20 entries after I batch up a few undead).

      The Reincarnation spell explicitly says to use that page for a random roll, so I use it for both languages and reincarnation that way. And also as a permitted list for polymorph targets (minus a quartet of high-HD types).