Japanese Honorifics as Emotive Definite Descriptions

Christopher Potts, Shigeto KawaharaProceedings From Semantics and Linguistic Theory 2004

Abstract

This paper provides a formal semantic framework for Japanese honorifics, arguing that they function as "expressives" rather than contributing to the basic truth-conditional meaning of a sentence. By treating honorifics as emotive definite descriptions, the authors explain why these markers are independent of the main message, cannot be negated or questioned, and fail to work with quantified subjects like "most professors."

The Semantic Challenge of Honorifics

In Japanese, honorifics are ubiquitous, yet linguists have historically struggled to define exactly what they contribute to the meaning of a sentence. While the syntax (how they are put together) is well-documented, the semantics (what they mean) remained vague.

The core challenge is that honorifics do not change the "facts" of a sentence. Saying "Sam laughed" and "The honorable Sam laughed" describe the same event. However, the latter adds a layer of speaker attitude. The authors address the lack of a formal "compositional" logic—a set of rules that explains how these social markers interact with the rest of the language's grammar.

Methods: Multidimensional Semantics

The authors employ a multidimensional semantic approach. They propose that a single sentence can carry two independent types of meaning simultaneously:

  1. Propositional Content: The basic "who did what" (Type $e$ for entities, Type $t$ for truth values).
  2. Expressive Content: The speaker’s emotional or social attitude (a new Type $\epsilon$).

To handle this, they introduce a novel composition rule that allows an honorific to "peek" at an argument (like a person's name), generate an expressive meaning, and then pass the original person's name up the grammatical tree unchanged.

The proposed composition rule (14) allows expressive content to be calculated alongside regular meaning without interfering with it.

They also quantify "emotion" using a scale of real numbers from -1 to 1. Positive honorifics fall between 0.5 and 1, while "antihonorifics" (markers of contempt) fall between -1 and -0.5. As a theoretical linguistics paper, this work relies on linguistic "minimal pairs"—carefully constructed Japanese sentences that differ only by the presence or absence of an honorific marker—to test how native speakers' intuitions change regarding truth and social appropriateness.

Key Insights and Ideas

The authors argue honorifics are defined by four traits: Nondisplaceability (they cannot be negated), Independence (social meaning is separate from factual meaning), Immediacy (the act of saying them is the act of showing respect), and Descriptive Ineffability (they are difficult to paraphrase in plain words).

Honorifics as Definite Descriptions

Just as "the cat" requires a specific cat to be familiar in the conversation, an honorific requires a "familiar" social relationship to exist in the context before it can be used felicitously. This leads to the observation that social superiority isn't always transitive.

Figure 1 illustrates that honorific relationships are not always transitive, complicating simple 'social hierarchy' models.

As shown in the diagram above, Shigeto might use honorifics with his father's students, and those students use them with his father, but Shigeto might use non-honorific forms with his father. This highlights why "social superiority" is an oversimplification of honorific meaning.

Key Results: The Quantifier Constraint

The paper successfully models why honorifics behave differently from regular adjectives. A major result is explaining the quantifier constraint. In Japanese, you generally cannot use a subject honorific with a quantifier like "most professors."

The authors' logic explains this: honorifics are like "non-restrictive modifiers" (similar to "Sam, who I respect..."). Because they function as definite descriptions, they need to point to a specific, unique individual. A quantifier like "most" refers to a set, not a unique individual, creating a logical clash that makes the sentence sound ungrammatical to native speakers.

The framework also accounts for "performative honorification," where the speaker isn't necessarily honoring the subject of the sentence, but rather speaking politely to the listener (the addressee). This includes variations of commands ranging from neutral to polite to contemptuous (antihonorifics).

Limitations and Future Work

While the model is robust, the authors acknowledge several areas for future research. They provide only a "sketch" of how this works in dynamic settings where context changes mid-conversation, suggesting integration with speech-act theory. Additionally, while they touch on Korean, they acknowledge that different languages divide the lexical class of honorifics differently, and their model might need tuning for systems like Thai. Finally, rare cases of "embedded expressives" in reported speech challenge the nondisplaceability rule and require further investigation.

Summary updated

Abstract

No abstract.
BibTeX
@inproceedings{potts2004japanese,
  author = {Christopher Potts and Shigeto Kawahara},
  title = {{Japanese Honorifics as Emotive Definite Descriptions}},
  booktitle = {Proceedings From Semantics and Linguistic Theory},
  year = {2004},
}
Figure