A person on screen says very little. The original performance is measured, almost restrained. A new-language voice makes the same person sound rough, exuberant, or threatening.
Perhaps the words are accurate. Perhaps the recording is excellent. Yet the character's place in the scene has shifted.
In dubbing, that shift can happen through qualities that are not usually treated as translation: pitch, vocal texture, pace, pronunciation, laughter, breaths, pauses, and the way a sentence ends. These do not simply decorate information. They participate in the person the audience hears.
This is where the question of embodiment becomes a question of representation.
Making a person legible
I use representation here for the process by which signs make a person, group, or world understandable in particular ways. It is not a transparent duplication of something already complete outside the work.
Image, dialogue, gesture, costume, narrative, and voice select and organize what becomes noticeable. A character can be heard as educated, hesitant, provincial, authoritative, intimate, or socially distant through the relations among those signs.
Such readings are neither universal nor reliable measurements of the person's actual attributes. They depend on conventions, context, and the listener. That is precisely why they require attention in performance.
An actor does not first create a neutral person and then add representation. Choices made in creating the performance already affect who that person appears to be.
For dubbing, the issue is particularly sharp. The new voice enters a character whose representation has already been partly established. It may reinforce that arrangement, transform it, or overwrite it with a familiar local type.
Representational judgment is therefore not just a final check for offensive content. It belongs to the initial question: whom are we performing?
Marked and unmarked
The terms marked and unmarked give me a useful way to ask what a work treats as a noticeable difference and what it allows to pass without explanation.
The terms have different uses across linguistics and cultural analysis. I use them here as a limited analytical aid, not as a single theory that can explain every social distinction.
To be unmarked is not to have no attributes. It is to occupy a position in which certain attributes do not need to be named.
Consider a context in which one person is called “a doctor” and another “a Black doctor,” or one is “a soldier” and another “a gay soldier.” The question is not whether those descriptions are always wrong. It is what the contrast is doing: why does an attribute become necessary to the description of one person and not the other?
These are examples of a possible asymmetry, not a claim that every culture, genre, or scene uses the same default. What is unmarked can change with the context.
Representation concerns more than the number of different people who appear. It also concerns who can appear as a person without qualification, and who is made to stand for an attribute before they are allowed to be an individual.
The supposedly ordinary person has a gender, ethnicity, history, and social position too. Those conditions may simply remain unannounced.
Dubbing redistributes the distinctions
The original performance may contain audible differences in region, class, generation, education, or relation to another person. Moving it into another language does not carry each difference automatically.
A regional accent may have no direct equivalent in the target language. A standard variety may carry different social meanings. A convention that signals refinement in one context may suggest distance or affectation in another.
Dubbing can preserve a distinction through a different device, make it less noticeable, or add a distinction that the original performance did not emphasize.
If several socially differentiated voices are all rendered in the target language's familiar standard register, the result may sound neutral. But neutrality is itself a choice: differences may have been removed, and the target culture's default substituted.
Conversely, imagine a Black male character whose restrained original delivery is replaced by a rougher, lower, more forceful voice because the casting or direction begins with a racialized type. This is a hypothetical example of the mechanism, not evidence about a named production. The additional vocal mark does not come from the character's action in the scene. It comes from a category imposed on him.
The dubbing process has done more than add emotion. It has changed how the person is classified and understood.
Two translations, not one
My Japanese draft began from a concern about how characters are received in Japan. People who occupy distinct positions within the original work's society may first be grouped under a broad category such as “foreigners,” then redistributed among familiar vocal types.
I want to keep this as a mechanism to investigate, rather than as a claim that Japanese viewers or performers uniformly respond this way.
The first movement compresses differences within the source world. The second makes the compressed figure legible through the target world's conventions. A character's ethnicity, class, region, or way of managing an interaction can be folded into a stock role before the original relations have been read.
This is not solved by saying that Japan lacks representation while another society understands it. That would substitute a national stereotype for the representational problem being examined.
The more useful questions concern particular works and decisions. Which distinction was removed? Which new one was added? What did the new voice make the audience hear? What evidence in the original performance supported that choice?
Those questions allow a comparison without assuming either culture is a neutral reference point.
Representational literacy
I call the ability to read these relations representational literacy. I do not mean learning the correct voice for each identity category.
A new catalogue—this is how a gay person sounds, this is how a Black person speaks, this is the voice appropriate to a transgender character—would reproduce the very problem it claims to solve.
The craft requires reading what the work makes salient, what it leaves unmarked, and how the character's speech arises from a particular body, situation, history, and relationship.
It also requires examining the performer's own defaults. What am I calling a natural voice? What conventions make this delivery seem ordinary? Am I reading a character, or selecting the sound I have learned to associate with a category?
Templates can be useful. They become hollow when their condensed conditions are forgotten. A familiar sound is an aid to recognition; it is not sufficient evidence that the character has been understood.
Greater expressive ability can amplify the wrong reading. A powerful performance of an imposed type may be less faithful to the person on screen than a quieter, more carefully grounded one.
Representation belongs to the craft because it helps determine the direction in which expressive skill is used.
Why acting ability does not settle casting
This leads to a different level of the problem: casting performers who share an identity or lived experience represented by a role.
One response is that actors play people other than themselves. That is true as a description of acting. It does not, by itself, answer who receives the opportunity to act.
The ability to perform a role, the history of how that group has been represented, and access to auditions and employment are distinct questions. A statement about ability cannot make the other two disappear.
Imagine a group whose stories are repeatedly performed by others while its own actors rarely gain access to the relevant roles. In that situation, asking who is cast concerns more than the quality of an individual audition. It concerns who is represented as an object of interest and who is permitted to participate as a performer.
This is a reason to examine the production conditions, not a universal instruction that every role must be allocated through identity matching.
Access without compulsory disclosure
Identity-conscious casting can expand opportunity. A rigid demand to prove identity can create another problem.
Sexuality, gender history, disability, and other aspects of lived experience are not always public information. A person may have reasons not to disclose them to employers, colleagues, family, or audiences.
If access to a role depends on public proof of belonging, a policy intended to include people can pressure them to reveal private information. The same risk appears when audiences treat a performance as evidence of an actor's personal identity.
Criticism of representation should not become an invitation to investigate a performer's private life. The questions are what the work represents, what opportunities the production provides, and what interpretive choices the performance makes.
Playing a queer character does not entitle an audience to demand a declaration from the actor. Nor should support for visibility become pressure to make every person's life visible.
This is an argument about production ethics, not legal advice or a description of one universal casting policy.
Keeping the levels separate
The choice is not exhausted by “only people with the identity may play the role” and “actors can play anyone, so identity is irrelevant.”
I want a framework that can ask several questions without turning them into one test.
Can actors from the represented group reach the audition and be taken seriously? If someone outside the group is cast, how does the production support reading the character and the history of the representation? Does the performance establish a particular person, or consume a set of identity tags? Does discussion of the work protect the performer's privacy?
These questions operate at different levels. None is answered merely by increasing vocal skill, and none requires reducing the actor to a statement of personal attributes.
The capacity to perform another person matters. So does the social arrangement in which that capacity is exercised.
Back to the voice
The question returns to the booth with greater specificity. What person is this new voice helping to establish? What relations does it preserve? What does it add that was not there?
It also extends to generated voices. A catalogue description or prompt can place a character within a social type before the scene has been read. Reproducing that type smoothly does not make the representational choice disappear.
For either a human or a synthetic voice, I want to ask whether the selected expression belongs to the person and action being represented.
Representational literacy is not a brake attached to acting from outside. It is one way to keep the ability to perform another person from becoming an efficient reproduction of a category.
The next note turns to localization: when translation, voice generation, and even the visible mouth can be changed, who decides what the character and the work must retain?
Method note: the examples are hypothetical unless otherwise identified. Markedness and representational literacy are used as analytical tools, not measurements of performers or national cultures. This note proposes questions for performance and production; it does not report a comparative survey of casting systems. The source-language discussion of national histories and training requires a separate, documented comparison before it can support general claims.
Previous: Dubbing as Re-embodiment · Next: What Should Survive Localization?