Sometimes the most important question in a sentence is not who performed the action, but whose perspective makes that action meaningful.
Imagine two men, Kofi and Daniel.
Kofi tells you something about an earlier event.
Later, you report the conversation to another person:
Kofi said that he left.
Who left?
Kofi?
Daniel?
Somebody else?
English does not necessarily tell us.
The pronoun he can refer to Kofi, but it can also refer to another man whose identity is understood from the context.
We normally resolve such ambiguity through the surrounding conversation.
We ask what has already been mentioned.
We consider who was present.
We interpret the speaker’s intentions.
Sometimes we simply ask:
Who do you mean?
Now imagine a language that has a special grammatical resource for referring to the person whose speech, thoughts or experiences are being reported.
A language in which the difference between two kinds of he can become visible in the pronoun itself.
That language is Ewe (Éʋegbe), spoken primarily in Ghana and Togo.
And the phenomenon is called logophoricity.
It leads us into a surprisingly deep question:
Can grammar represent another person’s point of view without turning that person’s words into a direct quotation?
1. The English sentence contains two possible stories
Consider:
Kofi said that he left.
We can interpret it in two ways.
Interpretation A
Kofi said:
“I left.”
The person who left is Kofi himself.
Interpretation B
Kofi said:
“He left.”
The person who left is someone other than Kofi.
Both interpretations can produce the same English sentence:
Kofi said that he left.
The words remain unchanged.
The underlying referential relationships do not.
That is the first distinction we need to understand.
A sentence is not merely a sequence of words.
It also contains relationships between expressions and the people, objects or events to which they refer.
2. Ewe makes the contrast visible
The following examples are documented in descriptions of Ewe grammar.
Kofi bé édzó.
Kofi said that he or she — someone other than Kofi — went away.
Compare:
Kofi bé yèdzó.
Kofi said that he — Kofi himself — went away.
The difference is small on the surface.
But its grammatical consequences are substantial.
The ordinary third-person pronoun é contrasts with the logophoric pronoun yè.
In this construction, the logophoric form identifies the reported speaker as the referent of the embedded pronoun.
The distinction is not expressed by changing the English translation from he to she.
Both versions may involve a male referent.
What changes is the relationship between the reported statement and the person whose statement it is.
This is why translating both sentences simply as:
Kofi said that he left
conceals an important part of the source-language structure.
3. What exactly is a logophoric pronoun?
The term logophoric comes from linguistic terminology concerned with reported discourse and perspective.
In Ewe, a logophoric pronoun can identify the individual whose speech, thoughts, feelings or other relevant conscious perspective is being represented.
That description matters because logophoricity is not limited to a person physically speaking aloud.
A language may report:
what somebody said,
what somebody thought,
what somebody knew,
what somebody felt,
or how somebody experienced a situation.
The central issue is the connection between an embedded expression and the individual whose perspective the report concerns.
Consider an English sentence:
Maria thought that she was in danger.
Who is she?
Maria?
Another woman?
Now imagine a grammatical distinction specifically associated with identifying the person whose thought is being reported.
The language is not merely identifying a woman.
It is identifying a woman in relation to the represented mental perspective.
That is the conceptual territory of logophoricity.
4. Why an ordinary pronoun is not enough to explain it
An ordinary third-person pronoun can identify someone.
He arrived.
She left.
They waited.
But a logophoric pronoun is not simply another interchangeable name for a participant.
Its interpretation is connected to a particular grammatical and discourse environment.
Compare the difference between these two instructions:
Find the man we are talking about.
And:
Find the person whose speech or thought this report represents.
The first is a general reference problem.
The second introduces a perspective relationship.
A logophoric form helps encode that second relationship.
It tells us something about the structure of the report, not only the identity of the participant.
5. A pronoun can point into someone else’s consciousness
Consider an ordinary statement:
Kofi left.
The narrator presents an event.
Now:
Kofi said that he left.
The narrator presents a report of something Kofi said.
The sentence contains two levels.
At the outer level, someone is reporting Kofi’s statement.
At the inner level, we have the content attributed to Kofi.
The narrator and the original speaker are not necessarily the same person.
This is where perspective becomes important.
The narrator is speaking now.
But the embedded proposition belongs to a different communicative or mental context.
A logophoric pronoun can help connect an expression inside that embedded proposition to the person whose perspective is being reported.
Grammar has crossed from reference into the organization of reported consciousness.
6. Direct speech solves the problem differently
English already possesses a powerful way of representing somebody else’s perspective.
Quotation.
Compare:
Kofi said, “I left.”
There is no ambiguity about the first-person pronoun inside the quotation.
The I belongs to Kofi’s original speech.
Now compare:
Kofi said that he left.
The pronoun becomes third person in the narrator’s report.
English has shifted from directly reproducing the original speaker’s words to describing their content.
A logophoric system offers another possibility.
It can preserve a special connection to the reported speaker while remaining within indirect discourse.
That does not mean the logophoric pronoun is simply an ordinary first-person pronoun disguised as a third-person form.
The relationship is more complicated.
7. Why “he means I” is a useful explanation — and a dangerous one
For a beginner, we might say:
Logophoric he = the original speaker’s I.
This helps explain:
Kofi said that he left.
But it is not a complete definition.
There are two different questions:
Who does the pronoun refer to?
And:
How does the person whose perspective is reported understand that reference?
Those questions may produce different answers.
To see why, we need a distinction from the philosophy and semantics of attitudes.
8. De se: thinking about oneself as oneself
Imagine a man looking into a mirror.
He sees someone with a stain on his jacket.
He thinks:
That man has a stain on his jacket.
He does not realize that the man in the mirror is himself.
Objectively, he is thinking about himself.
But subjectively, he is not thinking:
I have a stain on my jacket.
Now suppose he recognizes his reflection.
He thinks:
I have a stain on my jacket!
Something has changed.
The person referred to is the same.
But the way the thinker represents that person is different.
In semantic terminology, a thought explicitly involving oneself as oneself is often called a de se attitude.
The Latin expression means approximately about oneself.
The important distinction is not merely identity.
It is self-identification within the represented perspective.
9. De re: thinking about an individual without recognizing oneself
Return to the mirror.
The man thinks:
That man looks ridiculous.
Unknown to him, he is looking at himself.
We can truthfully report that he has a belief about himself.
But the belief is not necessarily self-conscious in the first-person sense.
This is the kind of situation used to explain a de re interpretation: an attitude concerning a particular individual, even when the thinker does not recognize that individual under the description used by the reporter.
The distinction is subtle.
Yet it is fundamental.
Being the object of your own thought is not always the same as knowing that you are thinking about yourself.
Now the problem becomes particularly interesting for Ewe.
10. Does a logophoric pronoun require de se interpretation?
It would be tempting to answer yes.
After all, a logophoric pronoun refers to the person whose perspective is being reported.
So perhaps it must always correspond to an internal I.
But linguistic research has shown that this conclusion is too strong for Ewe.
In her 2015 study, “The Interpretation of the Logophoric Pronoun in Ewe,” Hazel Pearson presents evidence that Ewe logophoric pronouns are not obligatorily interpreted de se.
That result matters enormously.
It means that the grammatical connection between a logophoric pronoun and an attitude holder does not automatically guarantee a first-person self-aware interpretation.
The pronoun can be connected to the individual whose attitude is being reported without every interpretation requiring that individual to recognize the referent as themselves.
This is precisely why logophoricity cannot be reduced to:
third-person form = hidden first-person thought.
The grammar is doing something more sophisticated.
11. The same person can exist in different mental representations
Imagine three descriptions:
The man in the mirror is late.
Kofi is late.
I am late.
Suppose all three refer to Kofi.
At the level of real-world identity, the referent is the same.
But from Kofi’s own perspective, the expressions may not be equivalent.
He might know that Kofi is his name.
He might fail to recognize himself in the mirror.
He might understand perfectly well that he himself is late.
Reference alone does not capture these differences.
To understand reports of belief, knowledge and experience, we must sometimes distinguish:
the person being referred to
from
the way that person is represented in someone’s mind.
Logophoricity brings these issues directly into grammar.
12. Grammar can track a perspective without reproducing a thought
This is one of the most important distinctions in the entire investigation.
Suppose a narrator reports somebody else’s experience.
The narrator may know more than the person being described.
The narrator may use expressions that the original thinker would never have used.
The narrator may identify a person whom the original thinker failed to recognize.
Yet the report can still be organized around that thinker’s perspective.
Therefore:
representing someone’s perspective does not necessarily mean reproducing their exact mental words.
That is a crucial insight for translation, narrative analysis and everyday communication.
It also explains why the relationship between logophoricity and self-conscious thought is more complex than it first appears.
13. Ewe has more than one logophoric form
The phenomenon is not restricted to the singular form yè.
Descriptions of Ewe also identify the plural logophoric form yèwó.
The contrast is primarily one of number.
This is significant because logophoricity is not simply an unusual singular pronoun.
It is part of a grammatical system for managing reference within reported perspectives.
The distinction also appears in environments involving second-person antecedents and in object positions.
For example, the Ewe grammar published by the University of Cologne gives:
èbé yèdzó.
You said that you went away.
And:
Ama bé mèkpɔ́ yè.
Ama said that I saw her — Ama.
These examples show that the system is not adequately described as:
a special word meaning “he himself.”
It is a more general strategy for identifying the relevant perspective holder across the structure of reported discourse.
14. Logophoricity is not an ordinary reflexive
English has reflexive pronouns:
myself
yourself
himself
herself
themselves
So we might ask whether Ewe yè is simply another form of himself.
But reflexivity and logophoricity are different mechanisms.
Consider:
Kofi washed himself.
The reflexive connects the object of washed to the subject within the relevant grammatical structure.
Now consider:
Kofi said that he left.
The problem is not whether the subject and object of left are identical.
There is no object of left in this example.
The problem is whether the subject of the reported event refers to the person whose statement is being reported.
That is a relationship across the architecture of reported speech.
The two phenomena can interact in natural languages, but they should not be collapsed into one category.
15. A special pronoun does not mean every pronoun works that way
Another important limitation concerns distribution.
In ordinary independent statements, Ewe does not simply replace all third-person pronouns with yè whenever a previously mentioned person is involved.
The logophoric form is associated with particular embedded and attitude-reporting environments.
Compare the conceptual difference:
Kofi arrived. He left.
And:
Kofi said that he left.
The first is a sequence of narrative statements.
The second contains an embedded report.
Logophoricity belongs to the special relationship established by the second kind of construction.
This prevents another misleading interpretation:
yè is not merely an unusually precise general-purpose pronoun.
Its grammatical environment is part of its meaning and distribution.
16. Why the source of a thought matters
Human communication contains several overlapping roles.
The person speaking now.
The person whose earlier words are reported.
The person whose thoughts are described.
The person experiencing an emotion.
The person from whose viewpoint a situation is presented.
Sometimes these roles coincide.
Sometimes they do not.
Imagine:
Anna said that Maria thought she was being followed.
Whose perspective governs she?
Anna’s?
Maria’s?
A third woman’s?
English requires us to reconstruct the intended relationships from syntax and context.
A language with dedicated logophoric resources can make some perspective relationships more explicit.
But the exact distribution depends on the language and construction.
Logophoricity is not a universal mechanism that automatically eliminates every possible ambiguity.
17. Perspective is not identical to grammatical subject
A useful distinction emerges.
A grammatical subject is a structural relation within a clause.
A perspective holder is the individual whose speech, thought, feeling or consciousness is being represented.
These roles frequently coincide in simple examples.
Kofi said…
Kofi is the subject of said.
Kofi is also the reported speaker.
But the conceptual roles should remain distinct.
In more complicated structures, the individual whose perspective matters cannot always be identified by a simplistic instruction such as:
Choose the nearest subject.
This is why serious analyses of logophoricity examine embedding, attitude predicates, reference and semantic interpretation rather than merely comparing two noun phrases.
18. Who owns the sentence?
This question sounds philosophical.
But it is linguistically useful.
Consider:
Kofi said that the decision was unfair.
The narrator utters the entire sentence.
Yet the evaluation unfair is attributed to Kofi.
The narrator may agree.
The narrator may disagree.
The narrator may simply report.
The sentence contains information whose source belongs to another person.
Now consider a pronoun inside the reported content.
Its interpretation may depend on that same source relationship.
Language therefore needs to manage not only:
who performed the event
but also:
whose representation of the event is being communicated.
Logophoricity is one grammatical solution to that problem.
19. A narrator can speak without becoming the centre of every perspective
This is particularly important in storytelling.
A narrator can describe a character’s actions.
Then report the character’s thoughts.
Then move into another character’s knowledge.
Then return to an external description.
The grammatical speaker remains the narrator.
But the perspective through which information is presented can change.
Literary analysis often discusses related phenomena through terms such as:
point of view
focalization
free indirect discourse
represented thought
These are not all equivalent to grammatical logophoricity.
But they address a related fundamental problem:
How can one person’s language represent another person’s experience?
Ewe gives us a grammatical entry point into that much larger question.
20. The difference between a voice and a viewpoint
Suppose a narrator says:
Kofi was frightened. The stranger was approaching.
The narrator is speaking.
But the description may be organized around Kofi’s experience.
Now:
Kofi thought the stranger was approaching.
The attribution becomes explicit.
And:
Kofi said that the stranger was approaching.
The narrator attributes the proposition to Kofi’s speech.
These are different relationships.
One involves narrative presentation.
One involves thought.
One involves reported speech.
Logophoric systems are particularly revealing because they show that languages can grammatically distinguish certain relationships between an expression and a represented consciousness.
The grammatical voice and the represented viewpoint need not be the same.
21. English is not incapable of representing another person’s perspective
This point is essential.
English can communicate all kinds of complex beliefs, mistaken identities and reported experiences.
Consider:
Kofi said that he himself had left.
Or:
According to Kofi, Kofi was the person who left.
Or:
Kofi claimed to have left.
These constructions are not interchangeable.
They have different syntactic and pragmatic properties.
But they demonstrate that English possesses resources for managing reference and perspective.
The contrast with Ewe is not:
one language understands perspectives and the other does not.
It is:
languages can distribute the grammatical work of perspective tracking differently.
That distinction protects us from linguistic determinism.
22. A translation can preserve identity while losing perspective marking
Return to the Ewe sentence:
Kofi bé yèdzó.
We translate:
Kofi said that he left.
The translation can be accurate in context.
The listener may understand that Kofi was the person who left.
But the English pronoun he does not, by itself, reproduce the dedicated logophoric distinction.
English may need context or additional wording to make the same relationship unmistakable.
The translation preserves the intended reference.
It does not necessarily preserve the grammatical mechanism through which the source language established that reference.
This is a recurring problem in comparative linguistics.
Preserving the answer is not always the same as preserving the question that the source grammar required the speaker to answer.
23. Amele tracked who continued acting. Ewe tracks whose perspective is represented.
In our earlier investigation of Amele and switch-reference, we examined how grammar can track whether the relevant subject remains the same across linked clauses.
Now Ewe takes us into another dimension.
Amele asks, in its canonical switch-reference pattern:
Does the next clause continue with the same subject?
Ewe logophoricity asks something different:
Does this pronoun refer to the individual whose speech, thought or experience is being reported?
These questions can both concern reference.
But they do not describe the same mechanism.
A participant can remain the same across events without being the holder of a reported perspective.
And a perspective holder can be identified through a logophoric pronoun without the construction being a switch-reference chain.
That difference is the reason these two studies belong next to each other without repeating each other.
24. The deeper problem is the architecture of consciousness in language
Human beings do not have direct access to one another’s private thoughts.
We hear statements.
Observe behaviour.
Receive reports.
Interpret expressions.
Construct models of what other people know, believe, fear or intend.
Language is one of the principal systems through which these models are communicated.
Consider:
She knows that he is coming.
She thinks that he is coming.
She hopes that he is coming.
She fears that he is coming.
The embedded proposition can remain similar.
But the relationship between the person and that proposition changes.
Knowledge.
Belief.
Hope.
Fear.
And when pronouns appear inside such reports, another question emerges:
Whose perspective determines their interpretation?
Logophoricity shows that this question can have grammatical consequences.
25. The mirror experiment reveals a hidden distinction
Return to the person who does not recognize himself in a mirror.
He sees a man.
He thinks the man looks tired.
He does not know the man is himself.
Now a narrator reports:
He thought that he looked tired.
English permits a reading in which the two pronouns refer to the same individual.
But the sentence does not automatically tell us whether the man recognized himself.
That is the difference between referential identity and first-person self-awareness.
The distinction matters for interpreting attitude reports.
And it is precisely the kind of distinction that prevents us from equating Ewe logophoricity with obligatory de se interpretation.
Pearson’s research demonstrates why the grammar of perspective cannot be reduced to the psychology of conscious self-recognition.
26. Why this matters for translation
Suppose a character mistakenly believes that a person in a photograph is someone else.
Later, the reader discovers that the person in the photograph was the character herself.
A translator must preserve the distinction between:
who the character actually referred to
and
who the character believed she was referring to.
Those are not necessarily the same representational facts.
If the translator resolves every pronoun according to objective identity, the character may suddenly appear to know something she did not know.
If the translator preserves only the character’s mistaken perspective, other referential relationships may become unclear.
This is why translating thought reports is not simply replacing pronouns.
It requires reconstructing the relationship between the narrator’s knowledge and the character’s knowledge.
Logophoricity makes this problem unusually visible.
27. Why this matters for language teaching
Students often learn reported speech through mechanical transformations:
I → he
my → his
am → was
will → would
Those transformations can be useful.
But they are not the whole phenomenon.
Consider:
John said, “I am right.”
Reported:
John said that he was right.
The mechanical exercise is straightforward.
But now introduce two men.
John told Peter that he was right.
Who was right?
John?
Peter?
The correct interpretation depends on more than a pronoun conversion table.
It depends on reference, syntax, context and communicative intention.
Ewe provides an excellent comparative demonstration that reported speech is fundamentally a problem of perspective management, not merely pronoun replacement.
28. Meaning first, grammatical form second
A more productive teaching sequence begins with the situation.
Who originally spoke?
Who is reporting the statement now?
Who is being referred to?
Does the original speaker recognize the referent as themselves?
What does the listener already know?
Only after these relationships are clear should we examine the grammatical forms used to express them.
This approach applies far beyond Ewe.
It is useful when teaching English reported speech, German indirect discourse, reflexive pronouns, reference tracking and narrative perspective.
The practical principle is simple:
Before changing a pronoun, establish whose perspective the sentence represents.
Otherwise the grammar exercise may be formally correct while the intended meaning is wrong.
29. A sentence can contain several centres at once
Consider a reported statement.
There is a speaker producing the report.
There may be a different person who originally made the statement.
There may be another person whose experience is being described.
There may be yet another individual referred to by a pronoun.
The sentence is therefore not necessarily organized around a single human centre.
It can contain multiple layers of reference and perspective.
A language may use ordinary pronouns, reflexives, complement structures, quotation, special logophoric forms or other devices to manage those layers.
Ewe demonstrates one particularly explicit solution.
It can mark the relationship between a pronoun and the relevant perspective holder through a dedicated pronominal form.
30. What Ewe does not prove
The existence of logophoric pronouns does not prove that Ewe speakers possess a special biological capacity for understanding other people’s minds.
It does not prove that English speakers have a weaker theory of mind.
It does not mean that all reported perspectives are perfectly unambiguous in Ewe.
Nor does it mean that every logophoric pronoun corresponds to a conscious first-person thought.
Those claims would go beyond the evidence.
What Ewe does show is sufficiently important without exaggeration:
a language can grammatically distinguish reference to the holder of a reported perspective from ordinary third-person reference.
And research on its interpretation shows that even this apparently clear distinction interacts with sophisticated questions about identity, belief and self-awareness.
31. Grammar does not simply tell us who someone is
It can also tell us how a person is related to a represented experience.
That is the central discovery.
A name identifies an individual.
An ordinary pronoun can recover an individual from context.
A reflexive can encode a particular grammatical relationship.
A logophoric pronoun can identify an individual through their role as the holder of a reported perspective.
These are different operations.
Treating all of them as mere substitutes for names conceals the architecture of reference.
32. A word can belong to one voice while pointing into another
The narrator says the sentence.
The character owns the reported thought.
The pronoun refers to an individual.
And the grammar may connect that individual to the character whose perspective is being represented.
This is why logophoricity is so revealing.
It shows that the structure of a sentence can reach beyond the immediate speaker.
Language can represent relationships between different centres of consciousness without simply reproducing direct speech.
But those relationships must be analysed carefully.
Perspective is not identical to reference.
Reference is not identical to self-recognition.
And self-recognition is not identical to grammatical person.
Each distinction adds another layer to the analysis.
Conclusion: whose “he” are we hearing?
We began with a deceptively simple sentence:
Kofi said that he left.
English allows us to interpret the pronoun through context.
Ewe offers a dedicated grammatical distinction involving the person whose speech or thought is being reported.
At first, this looks like a convenient way to eliminate ambiguity.
But the deeper investigation takes us much further.
A person can refer to themselves without recognizing themselves.
A narrator can report another person’s thought without reproducing its exact internal wording.
A pronoun can be grammatically connected to a perspective holder without requiring every interpretation to be consciously first-person.
And two languages can communicate the same event while organizing its reference relationships differently.
Ewe therefore gives us more than an unusual pronoun.
It gives us a way to examine how language represents the boundary between the speaker’s perspective and someone else’s.
That boundary is present in every conversation involving reported speech, belief, memory, knowledge or emotion.
Some languages leave more of the work to syntax and context.
Others possess specialized grammatical resources.
The underlying human problem remains universal.
We speak from our own position.
But we constantly talk about what other people say, know, believe and experience.
And grammar sometimes helps us cross that boundary.
The most revealing pronoun is not always the one that tells us who someone is. Sometimes it is the one that tells us whose world we have entered.
Further reading and linguistic sources
The Ewe examples and grammatical descriptions in this article are based on documented linguistic sources rather than reconstructed or invented Ewe sentences.
- Basic Ewe Grammar — University of Cologne. See the section on logophoric pronouns.
- Hazel Pearson, The Interpretation of the Logophoric Pronoun in Ewe (2015). An open-access investigation of logophoric reference and the distinction between de se and de re interpretation.
- Decomposing Logophoric Pronouns: A Presuppositional Account of Logophoric Dependencies. A more recent formal-semantic treatment of Ewe logophoricity.
About the author and individual language learning
Tymur Levitin is the founder and director of Levitin Language School, an international online school providing individual language education.
The school’s approach connects grammatical structures with meaning, real communication, context and the decisions speakers make while using language.
These comparative linguistic investigations are not presented as claims that every language discussed is available as a taught course. Instead, they contribute to a broader understanding of how language works — knowledge that can enrich the study of English, German and other languages.
Individual online lessons are available for learners who want to work on real communication, grammar, vocabulary, comprehension, professional language or academic goals.
You can also explore Language Learnings for the school’s international learning opportunities.
For individual lessons or enquiries, contact Tymur Levitin directly on Telegram.
Tymur Levitin
Founder & Director, Levitin Language School
Language Teacher, Translator and Author
© Tymur Levitin / Levitin Language School. All rights reserved.
