Status: Established concept
Terminological origin: Established terminology in logic, rhetoric, linguistics, semantics and philosophy of language, subsequently developed across cognitive science, law and computational linguistics.
Relations: related to → Semantic Ambiguity · Lexical Ambiguity · Syntactic Ambiguity · Polysemy · Vagueness · Underspecification · Context Sensitivity · Indeterminacy · Incoherence · Interpretation · Disambiguation · Provenance · AI Alignment · Semantic Engineering · QualComputing
Concise definition
Ambiguity is a property of an expression or sign that supports more than one legitimate interpretation within a relevant linguistic or representational system.
In technical linguistic and philosophical usage, ambiguity should not be equated with every situation in which meaning is uncertain, incomplete or context-dependent. An expression may leave information unspecified without being ambiguous, and a vague expression may have uncertain boundaries of application without possessing several distinct meanings.
Ambiguity therefore concerns a plurality of legitimate interpretations or mappings, rather than merely an absence of precision.
Conceptual scope
Ambiguity occurs at different levels of linguistic and representational analysis.
A lexical item may support more than one interpretation. A syntactic structure may admit more than one structural analysis. Scope relations may generate alternative semantic readings. At other levels, interpretation may depend upon discourse, pragmatic or contextual information.
The common feature is not simply uncertainty, but the availability of more than one legitimate interpretation associated with an expression or representation.
The precise characterization of ambiguity nevertheless depends partly on the theoretical framework adopted. Kennedy, for example, characterizes ambiguity in terms of uncertainty in mappings between levels of representation having different structural characteristics, distinguishing it from vagueness, which concerns uncertainty associated with meaning itself.
Ambiguity should therefore be understood as a family of related interpretive phenomena rather than as a single homogeneous mechanism.
Key distinctions
Ambiguity vs. vagueness
Ambiguity and vagueness are distinct forms of interpretive uncertainty.
Ambiguity involves alternative interpretations or representational mappings. Vagueness typically concerns indeterminate boundaries of application and borderline cases.
An expression may therefore be ambiguous without being vague, vague without being ambiguous, or potentially exhibit both properties.
Ambiguity vs. underspecification
Underspecification occurs when a representation leaves some information unresolved without necessarily encoding several distinct meanings.
In formal and computational semantics, semantic underspecification also designates techniques that deliberately represent several possible readings without immediately resolving the differences between them. Underspecification can therefore function as a method for representing ambiguity without requiring premature disambiguation.
Ambiguity and underspecification must consequently not be treated as interchangeable concepts.
Ambiguity vs. context sensitivity
Context-sensitive expressions vary in content or reference according to features of their context of use. Such variation does not in itself establish ambiguity.
Context may, however, provide information through which an ambiguity is resolved. The fact that context contributes to disambiguation should therefore not be confused with context sensitivity as a linguistic property.
Ambiguity vs. polysemy
Polysemy concerns a lexical item possessing multiple distinct but related interpretations. It is commonly treated as a form of lexical ambiguity, but its internal structure is complex and heterogeneous.
Recent interdisciplinary work indicates that relations between polysemous senses may form a continuum ranging from near identity of sense to contrasts approaching homonymy. This makes simple binary distinctions between “one meaning” and “several meanings” insufficient for many lexical phenomena.
Ambiguity vs. indeterminacy
Ambiguity is one possible source of interpretive indeterminacy, but indeterminacy is broader than ambiguity.
An interpretation may remain indeterminate because of vagueness, underspecification, insufficient contextual information or other factors without the expression necessarily possessing several distinct meanings.
This distinction is particularly important when moving between linguistics, artificial intelligence and law, where ambiguity is sometimes used more broadly than in technical semantic analysis.
Ambiguity vs. incoherence
Ambiguity and incoherence concern different dimensions of interpretation. Ambiguity concerns the availability of more than one legitimate interpretation of an expression or representation. Incoherence concerns difficulties arising from the relations among elements within a larger representational or interpretive structure.
An ambiguous expression may support several internally coherent interpretations and is therefore not incoherent merely by being ambiguous. Conversely, a representational structure may be incoherent even when its individual elements are unambiguous.
The two phenomena may nevertheless interact. An unresolved ambiguity may be interpreted differently across successive contexts or representations, potentially contributing to downstream incoherence. Conversely, an incoherent context may weaken the constraints required to stabilize interpretation, thereby increasing interpretive indeterminacy without necessarily creating ambiguity in the strict linguistic sense.
Ambiguity and incoherence should therefore be distinguished analytically while their possible interactions remain open to investigation.
Across disciplines
Linguistics and semantics
Linguistics distinguishes several forms of ambiguity, particularly lexical and structural ambiguity, while semantic theory examines how alternative readings can be represented and constrained.
Formal approaches to underspecification demonstrate that semantic processing need not always require immediate selection of a single reading. Reyle’s work, for example, represents scope ambiguities through underspecified structures, while later work reviews broader families of underspecification formalisms.
Logic and philosophy of language
Ambiguity has long been relevant to logic because changes in interpretation can generate equivocation. An argument may appear formally acceptable while relying upon different interpretations of the same expression at different points.
Philosophy of language consequently examines both the nature of ambiguity and the conditions under which alternative interpretations can be identified and resolved.
Cognitive science and psycholinguistics
Ambiguity provides a means of investigating how human interpreters activate, represent and select among possible meanings.
Research on polysemy indicates that human processing cannot always be adequately represented by a simple model in which discrete predefined senses are selected independently of contextual and representational factors.
Artificial intelligence and natural language processing
Ambiguity remains a significant problem for natural language processing and large language models.
Contemporary research examines ambiguity detection, clarification and disambiguation in conversational systems and identifies unresolved problems in increasingly agentic environments.
The problem is not limited to whether an AI system can select a plausible interpretation. It also concerns when clarification is required, what contextual evidence supports interpretation and how uncertainty should be handled when several readings remain possible.
Law and regulation
Legal interpretation distinguishes ambiguity from vagueness and other forms of linguistic indeterminacy, although ordinary legal discourse does not always preserve these distinctions consistently.
The legal context introduces an additional problem: the relevant question may concern not only which contextual information could resolve an interpretation, but which interpretive resources are legally legitimate or institutionally authorized.
Ambiguity can therefore acquire practical significance when alternative interpretations generate different rights, obligations, standards of compliance or allocations of responsibility.
Conceptual Issues / Points of Debate
Ambiguity appears simple when defined as the availability of multiple interpretations, but several theoretical problems remain.
First, what bears the ambiguity? Depending on the theoretical framework, ambiguity may be attributed to lexical items, syntactic structures, sentences, utterances or other representational structures.
Second, what counts as a distinct interpretation? Two apparent readings may correspond to genuinely different meanings, related polysemous senses, contextual specifications of a more general meaning, or pragmatic elaborations.
Third, must alternative interpretations be explicitly represented? Formal underspecification demonstrates that multiple readings can sometimes be represented without immediate resolution, while theories of language processing differ over whether interpreters necessarily construct explicit alternative representations.
Fourth, ambiguity is not intrinsically defective. Natural language routinely tolerates and exploits ambiguity, and contextual interpretation frequently resolves it without difficulty.
The relevant problem is therefore not simply whether ambiguity exists, but when unresolved or inadequately resolved ambiguity becomes consequential.
BSI perspective
From a BSI perspective, ambiguity becomes particularly relevant when alternative interpretations may lead to materially different downstream representations, inferences, decisions, obligations or actions.
Not every ambiguity therefore requires preservation, explicit resolution or traceability. In many contexts, interpretive alternatives can be resolved without significant consequences. Conversely, preserving every theoretically possible interpretation would itself generate unnecessary representational and computational complexity.
The relevant question is whether a distinction between interpretations matters to the reliability, legitimacy or consequences of the process in which the expression is used.
This becomes particularly important when meaning passes through successive human and artificial agents, disciplines or representational systems.
Existing research already addresses important components of this problem through ambiguity resolution, semantic underspecification, natural-language processing and related approaches. BSI therefore does not present the preservation or resolution of ambiguity as a new problem.
Instead, BSI identifies a broader research question concerning cases in which changes in interpretive possibilities may materially affect downstream processes:
Which interpretive distinctions matter, what happens to them across successive representations and transformations, and under what conditions should those changes remain traceable?
This question intersects with provenance, AI alignment and incoherence.
A system may produce a determinate output from an input whose interpretation was not itself determinate. Computational determinacy does not therefore establish that the preceding interpretive problem was adequately resolved.
Similarly, divergent resolutions of an earlier ambiguity may generate representations that subsequently become difficult to reconcile. Ambiguity may therefore contribute to incoherence without being itself a form of incoherence. Conversely, incoherence may destabilize the contextual relations through which interpretation is constrained, producing indeterminacy without necessarily producing ambiguity in the strict sense.
At the same time, semantic traceability should not become an objective in itself. From a QualComputing perspective, greater preservation of information, distinctions or provenance is not automatically equivalent to greater quality.
Converge where justified. Preserve difference where it matters.
The possible systematic tracing of materially significant changes in interpretive possibilities remains an open BSI research question, rather than an established BSI concept or taxonomy.
Sources
Sennet, A. (2021). “Ambiguity.” The Stanford Encyclopedia of Philosophy.
Source contribution: Provides a broad philosophical and linguistic treatment of ambiguity and distinguishes it from vagueness, context sensitivity, underspecification and generality. It also examines ambiguity in relation to logic, cognition, communication and law.
BSI relevance: Provides the principal terminological safeguard against treating all forms of semantic uncertainty as ambiguity.
Kennedy, C. (2011). “Ambiguity and Vagueness: An Overview.” In Maienborn, C., von Heusinger, K. & Portner, P. (eds.), Semantics: An International Handbook of Natural Language Meaning, Vol. 1, pp. 507–535.
Source contribution: Provides a systematic distinction between ambiguity and vagueness and characterizes ambiguity in relation to mappings between representational levels.
BSI relevance: Demonstrates that relations between representational levels are already part of established ambiguity theory. BSI should therefore not claim representational transition itself as a conceptual innovation.
Reyle, U. (1993). “Dealing with Ambiguities by Underspecification: Construction, Representation and Deduction.” Journal of Semantics, 10(2), 123–179.
Source contribution: Develops a formal approach in which scope ambiguities are represented through underspecified structures rather than requiring immediate enumeration and resolution of fully specified readings.
BSI relevance: Establishes an important antecedent for preserving interpretive alternatives without premature resolution.
Egg, M. (2010). “Semantic Underspecification.” Language and Linguistics Compass, 4(3), 166–181.
Source contribution: Reviews semantic underspecification as a technique for representing several readings of an ambiguous expression within a single representation while deliberately leaving their differences unresolved.
BSI relevance: Prevents BSI from presenting preservation of interpretive plurality as a novel idea and redirects attention toward the conditions under which changes in such plurality become consequential and potentially traceable.
Haber, J., & Poesio, M. (2024). “Polysemy. Evidence from Linguistics, Behavioral Science, and Contextualized Language Models.” Computational Linguistics, 50(1), 351–417..
Source contribution: Provides an interdisciplinary review of polysemy across linguistics, behavioral science and contextualized language models and demonstrates the heterogeneity of relations among lexical senses.
BSI relevance: Guards against treating multiplicity of interpretation as a homogeneous phenomenon and connects established lexical-semantic research with contemporary language models.
Solan, L. M. (2018). “The Interpretation of Legal Language.” Annual Review of Linguistics, 4, 337–355.
Source contribution: Examines ambiguity, vagueness and other forms of linguistic indeterminacy in authoritative legal texts, emphasizing that the context used for interpretation may itself be constrained by legal principles.
BSI relevance: Supports the distinction between a linguistically possible interpretation and an interpretation whose selection is legitimate within a normative or institutional framework.
Tanjim, M. M., In, Y., Chen, X., Bursztyn, V., Rossi, R. A., Kim, S., Ren, G.-J., Muppala, V., Jiang, S., Kim, Y., & Park, C. (2025). “Disambiguation in Conversational Question Answering in the Era of LLMs and Agents: A Survey.” Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 9537-9590. Association for Computational Linguistics.
Source contribution: Reviews ambiguity detection, clarification and resolution in LLM-based conversational question answering and identifies continuing research problems, including in agentic environments.
BSI relevance: Establishes that ambiguity remains operationally relevant in contemporary AI while preventing BSI from presenting ambiguity management in LLMs as a novel research problem.
Related Concepts
Semantic Ambiguity
Ambiguity arising at the level of semantic interpretation. Its precise boundaries depend upon the theoretical framework used.
Lexical Ambiguity
Multiplicity of interpretation associated with a lexical item.
Syntactic Ambiguity
Multiplicity of structural analyses associated with an expression or sentence.
Polysemy
Multiple related interpretations associated with a lexical item.
Vagueness
Indeterminacy associated particularly with boundaries or conditions of application rather than discrete alternative interpretations.
Underspecification
Representation in which relevant information or distinctions remain unresolved; in semantic theory, it can provide a means of representing multiple readings without premature disambiguation.
Context Sensitivity
Systematic variation in content or reference according to contextual parameters.
Indeterminacy
A broader condition in which interpretation is not fully determined; ambiguity is one possible source of indeterminacy.
Incoherence
A condition concerning problematic relations among elements within a larger representational or interpretive structure. Ambiguity and incoherence are analytically distinct, although unresolved or divergently resolved ambiguity may contribute to downstream incoherence.
Disambiguation
The process through which alternative interpretations are distinguished or one interpretation is selected.
Provenance
The traceable history of information, representations, agents and transformations. Its relationship to interpretive change remains an active area of BSI investigation.
AI Alignment
The problem of ensuring that AI behaviour, objectives and representations remain appropriately related to intended human, normative or institutional requirements. Ambiguity becomes alignment-relevant when materially different interpretations can generate different objectives, representations or actions.
Research status
Established concept; BSI extension under investigation.
The general concept of ambiguity and its principal distinctions are well established. BSI does not propose a redefinition of ambiguity.
Current BSI research investigates a narrower question at the intersection of ambiguity, incoherence, provenance, semantic transformation and AI alignment: whether, when and how materially significant changes in interpretive possibilities should remain traceable across successive human and artificial representations.
No claim of conceptual novelty is made at this stage.
Last revised: September 2026