A language can draw a boundary differently without making the reality on either side disappear.
Imagine pointing to your hand and asking someone:
What is this?
For an English speaker, the answer seems almost embarrassingly obvious.
A hand.
The category feels natural because the anatomy feels natural. The wrist seems to provide a visible boundary. Above it is the arm; below it is the hand. Fingers belong to the hand. The whole structure appears to be sitting there in the human body, waiting for language to give it a name.
So it is very easy to make an unconscious assumption:
the body already contains the category “hand,” and languages merely attach different words to it.
Now travel — linguistically, at least — to Rossel Island in Papua New Guinea.
There we encounter Yélî Dnye, an extraordinary language whose ordinary body-part vocabulary does something that initially looks almost impossible from an English perspective.
In its ordinary lexical system, there is no straightforward designation corresponding to English hand.
Nor is there an ordinary equivalent dividing off foot in the familiar English way.
And there is no single ordinary term corresponding to the whole English category leg.
That sounds like the beginning of a spectacular claim about human cognition.
It is not.
In fact, Yélî Dnye gives us something much more valuable:
a way to understand why such spectacular claims are often wrong.
1. The human body is universal. Its lexical map is not.
Humans on Rossel Island do not have different anatomy.
Neither do English speakers, Finnish speakers, Ukrainian speakers or Basque speakers possess fundamentally different sets of limbs.
Our bodies provide an unusually useful test case for linguistic categorization precisely because the physical object being categorized is so similar.
A tree can vary enormously.
A landscape can vary.
Political systems, foods, technologies and social institutions can vary.
But the basic external architecture of the human body gives languages a remarkably similar physical reality to work with.
And yet languages do not necessarily cut that reality into the same named pieces.
This distinction is fundamental:
physical structure ≠ lexical structure.
2. What does Yélî Dnye actually do?
Stephen C. Levinson’s detailed study of Yélî Dnye body-part terminology describes a system considerably more complex than a simple list of words.
Visible parts of the body participate in several kinds of semantic organization.
There is a partonomic subsystem: a hierarchy of parts and subparts.
There are terms for particular surfaces.
There are collections of surface features.
There are also taxonomic relations among certain terms.
Within the major partonomic organization, the body is divided into nine principal regions:
- the head,
- the neck,
- two upper limbs,
- the trunk,
- two upper legs,
- two lower legs.
Already, something should feel unfamiliar.
English does not normally present the body to us through precisely that lexical architecture.
And this is where the temptation begins.
3. “They have no word for hand”
This sentence is both interesting and dangerous.
It can mean something perfectly defensible:
The ordinary lexical system does not contain a term whose semantic extension corresponds neatly to English hand.
But people often hear a much larger claim:
Speakers do not possess the concept HAND.
Those are not the same statement.
Not remotely.
A lexical gap is not automatically a conceptual gap.
A language can lack a conventional single lexical item corresponding exactly to another language’s category while its speakers remain perfectly capable of perceiving, referring to, reasoning about or constructing that category when necessary.
Yélî Dnye gives us unusually powerful evidence for precisely this distinction.
4. The taboo vocabulary changes the experiment
Here the story becomes much more interesting.
Yélî Dnye has had special registers associated with social taboos and avoidance.
And in an alternative taboo vocabulary, distinctions emerge that demonstrate that the relevant bodily segmentation is conceptually available.
This matters enormously.
If speakers can operate with a different lexical division under another communicative condition, then the absence of the ordinary word cannot reasonably be treated as evidence that they are incapable of conceptualizing the relevant body region.
The body did not change.
The speakers did not suddenly acquire a new perceptual organ.
What changed was the lexical system being used.
That gives us something close to a natural linguistic experiment:
**same speakers
- same human anatomy
- different lexical register
= different available lexical segmentation.**
And suddenly the simplistic equation
NO WORD = NO CONCEPT
collapses.
5. A word is not a concept with a label attached
This is one of the most persistent illusions in language learning.
We often imagine the relationship like this:
OBJECT → CONCEPT → WORD
There is a thing.
Humans form one obvious concept of that thing.
Each language supplies its own label.
So:
hand = X = Y = Z
and dictionaries simply tell us which label belongs to which language.
But actual lexical semantics is rarely that clean.
A language may group together what another language separates.
It may separate what another language groups together.
It may lexicalize one level of a hierarchy while leaving another level without a conventional basic term.
It may organize the same domain differently in ordinary vocabulary and in a specialized register.
The translation problem is therefore not always:
What is their word for our thing?
Sometimes the more accurate question is:
Does their lexical system contain the same category boundary at all?
6. English “hand” feels anatomical because English has trained us well
Consider how natural the category hand feels.
It seems to belong to biology rather than language.
But notice what has happened.
We have spent our entire lives hearing:
hand
and:
arm
as separate lexical units.
We learn expressions involving them.
We receive instructions using them.
We describe pain through them.
Doctors, teachers, parents, clothing, gestures and everyday conversation continually reinforce the distinction.
Eventually the lexical boundary becomes psychologically effortless.
That does not mean English invented the wrist.
It means that English makes one particular segmentation of the limb lexically routine.
Those are different claims.
7. The wrist exists even when “hand” does not
This distinction protects us from an equally silly opposite conclusion.
If lexical categories are language-specific, should we say that reality has no structure and language arbitrarily invents everything?
No.
The human body contains bones, joints, muscles, surfaces and functional divisions.
The wrist is not a linguistic hallucination.
Fingers really differ anatomically from the forearm.
But physical reality usually offers many possible boundaries simultaneously.
Language does not create all of them.
It selects some of them for conventional lexical treatment.
A joint can motivate a boundary.
Function can motivate another.
Shape can motivate another.
Cultural practice can make another useful.
The existence of a physical boundary therefore does not guarantee that every language will lexicalize exactly that boundary at exactly the same hierarchical level.
8. “Leg” creates the same problem in reverse
English itself can help us see the trap.
What exactly is a leg?
In everyday English, the word commonly refers to the entire lower limb.
But in some anatomical contexts, leg can be used more narrowly for the region between the knee and ankle.
So even inside English, a familiar body word can operate at different levels depending on register and context.
Now compare that with a system in which upper and lower portions of the lower limb receive lexical recognition without an ordinary superordinate corresponding neatly to the everyday English leg.
Which language has found the “real” leg?
The question is badly formed.
The anatomy supplies structures.
Languages construct lexical categories over those structures.
9. Even “face” need not be one simple part
English makes face feel like another obvious unit.
But semantic systems do not have to represent every body region through the same kind of part-whole relationship.
Levinson’s analysis distinguishes, among other things, genuine partonomic relations from collections of surface features.
That is important because a semantic field does not need to be internally uniform.
Some expressions can identify parts.
Others can identify surfaces.
Others can collect features into a region.
Others can classify subtypes.
Our dictionary habit encourages us to flatten all of them into:
BODY PART = WORD
But the lexical architecture can be much richer.
10. The body is not merely cut into pieces
Imagine drawing a human figure and asking speakers of different languages to mark the extension of every body-part word.
At first we might expect merely different labels.
Instead, several things could vary:
where a boundary is drawn;
whether a region has its own conventional name;
whether two regions belong under one larger named unit;
whether the larger unit itself is lexicalized;
whether the expression refers to a part, surface or collection of features;
whether another register reorganizes the same domain.
The interesting object of study is therefore no longer the vocabulary list.
It is the semantic map.
11. Translation hides the map
Suppose a Yélî Dnye expression is translated into English as something involving the arm, hand, leg or foot.
A successful translation may tell an English reader exactly which physical region is relevant.
Communicatively, the translation works.
Semantically, however, something may disappear.
English has reconstructed the referent using English lexical boundaries.
This is a recurring problem in translation.
The translation preserves what happened in the world while quietly rebuilding the way the source language categorized it.
We have already seen related phenomena elsewhere.
In Fijian possession, English my can conceal grammatical distinctions concerning the relationship between possessor and possessed.
In Basque participant alignment, familiar English subject–object expectations are not enough to describe the case system.
And in Plains Cree, translation can preserve who sees whom while hiding the direct–inverse architecture through which the verb organizes those participants.
Yélî Dnye gives us another version of the same larger lesson.
Translation can preserve the referent while replacing the categorization.
12. Does vocabulary determine perception?
This is where popular discussions of language often become too dramatic.
If Yélî Dnye lacks an ordinary equivalent of hand, does that mean speakers literally see the human body differently?
The evidence does not justify that conclusion.
They can see the same anatomy.
They can interact with the same functional regions.
And, crucially, alternative vocabulary demonstrates that the conceptual distinction can be available even when the ordinary lexical system does not package it as English does.
So we should resist both extremes.
The first extreme says:
Languages merely put different labels on identical concepts.
Too simple.
The second says:
If a language lacks a word, its speakers lack the concept.
Also too simple.
The interesting territory lies between them.
Language can make particular distinctions habitually available, conventionally packaged and communicatively economical without imprisoning human thought inside its dictionary.
13. What lexicalization actually changes
Suppose English had no word hand.
Every time we wanted to identify that region, perhaps we would need to say something like:
the end of the upper limb below the wrist
or use some context-dependent expression.
Could we still conceptualize the region?
Of course.
But talking about it would be less economical.
Once a conventional word exists, a complex region becomes easily reusable as a discourse object.
We can say:
my hand
your hand
raise your hand
hand injury
hand movement
by hand
and build metaphorical extensions around the category.
Lexicalization therefore matters.
But it matters without requiring the absurd conclusion that unnamed things cannot be thought.
A word can make a distinction cheap to communicate without making it possible to perceive.
14. This also changes how vocabulary should be taught
Traditional vocabulary learning encourages pairs:
hand — translation
foot — translation
leg — translation
The method works reasonably well when the categories overlap closely enough.
But it trains the learner to assume that translation equivalence is the default structure of the world.
A better question is:
What region does this word actually cover?
Then:
What larger category contains it?
What smaller categories belong inside it?
Is the relationship partonomic, taxonomic or something else?
Does the language have a conventional term at the same level as mine?
Does register change the system?
Now vocabulary becomes semantic analysis rather than flash-card substitution.
This is precisely the difference between memorizing a word and understanding the system that made the word useful.
15. One body can support several linguistic anatomies
The expression linguistic anatomy must be handled carefully.
We are not claiming that languages rewrite human biology.
They do not.
What they can reorganize is the lexical representation placed over that biology.
Think of a map.
A physical territory remains the same.
One map shows countries.
Another shows elevation.
Another shows rainfall.
Another shows railways.
The maps are not interchangeable.
But neither does any map create the territory.
The human body is the territory.
Body-part vocabularies are different maps laid across it.
Yélî Dnye is especially valuable because its map makes some boundaries visible that English speakers rarely notice as linguistic choices at all.
16. And then the taboo vocabulary gives us a second map
This may be the deepest part of the entire case.
If we compared only English and ordinary Yélî Dnye vocabulary, we might still be tempted to tell a dramatic story about two languages producing two different conceptual worlds.
But the existence of alternative vocabulary inside Yélî Dnye itself complicates that story.
Now one language community can provide more than one lexical mapping of the same anatomy.
That forces us to separate at least four things:
physical reality
conceptual possibility
ordinary lexicalization
register-specific lexicalization
Once these layers are separated, many bad arguments about language and thought become much easier to detect.
17. No word does not mean no thought
Consider concepts for which English has no convenient single word.
English speakers routinely understand experiences that require phrases rather than individual lexical items.
We do not conclude:
English speakers cannot experience this because English requires five words to describe it.
Yet when an unfamiliar language lacks one of our familiar words, people suddenly become willing to make exactly that inference.
That is an asymmetry worth noticing.
Our own lexical gaps feel like vocabulary.
Other people’s lexical gaps are too easily romanticized into psychology.
Yélî Dnye gives us an excellent antidote.
18. But words are not irrelevant either
After rejecting linguistic determinism, we should not swing to the opposite extreme and declare vocabulary unimportant.
Lexical systems reveal recurring choices.
They tell us which distinctions have become conventional.
They show which regions can be referred to economically.
They create networks of part-whole relations, metaphorical possibilities and conventional associations.
A language that routinely lexicalizes one segmentation and a language that routinely lexicalizes another are not doing exactly the same linguistic work.
The careful claim is therefore neither:
words determine thought
nor:
words are arbitrary labels with no cognitive significance.
It is:
lexical systems are conventional ways of organizing distinctions for communication, and those systems can differ even when the underlying physical and conceptual possibilities overlap.
That is already a profound fact about human language.
19. The most interesting question is no longer “What is the word?”
When encountering a new language, learners often ask:
What is the word for X?
Sometimes that is exactly the right question.
Sometimes it is not.
The deeper sequence is:
Does this language lexicalize X as one conventional category?
If yes:
Does its boundary match mine?
If not:
How does the language divide the domain instead?
And finally:
What can speakers still conceptualize even when ordinary vocabulary does not provide a one-word category?
That final question prevents semantic typology from becoming mythology.
Conclusion: the hand was there all along
The striking fact about Yélî Dnye is not that its speakers somehow live without hands.
Nor is it that English speakers have discovered the objectively correct way to divide the human body.
The striking fact is that a body whose physical structure seems completely obvious can support different lexical organizations.
English makes hand, foot and leg feel like inevitable pieces of reality.
Yélî Dnye reminds us that some of that inevitability belongs not to anatomy but to the lexical map through which we have learned to describe anatomy.
And its alternative vocabulary gives us an even more important lesson:
when a category is absent from ordinary vocabulary, the concept itself has not necessarily disappeared.
The difference between those two statements is the difference between linguistic analysis and linguistic folklore.
A missing word is evidence about a language. It is not, by itself, evidence of a missing thought.
Tymur Levitin
Founder & Director, Levitin Language School
Certified Translator and Language Teacher
© Tymur Levitin / Levitin Language School. All rights reserved.
