The Structure of a Linguistics PhD Proposal
How to turn a linguistic intuition into a research problem, a theory, and an original contribution
A PhD proposal in linguistics is often mistaken for a formal description of what a researcher intends to study. It is not. At its highest level, a proposal is a demonstration that a particular linguistic phenomenon contains an unresolved problem, that existing theories do not adequately explain it, and that the proposed research has the empirical and theoretical resources to resolve it.
This distinction is especially important in syntax and morphology. A student may begin with an apparently promising topic, agreement in Urdu, case in Saraiki, pronoun interpretation in English, word order in Punjabi, or inflectional morphology in a lesser-described language. But a topic is not yet a research problem. The intellectual journey begins when the researcher can show that the phenomenon creates a genuine challenge for an existing account of grammar.
The most useful way to think about a proposal, therefore, is not as a collection of sections but as a chain of reasoning:
phenomenon → generalization → problem → theoretical gap → research question → analysis → evidence → prediction → contribution.
If that chain is broken, the proposal becomes descriptive. If it is coherent, the proposal begins to look doctoral. Consider a simple example. Suppose a researcher notices that verbal agreement in a language does not always correspond to the argument that appears to be the subject. The beginner asks: Why does this language behave differently? The more advanced researcher asks: What determines agreement in these constructions? The doctoral researcher asks a more consequential question: What does this pattern tell us about the structural conditions under which agreement is established? That final question changes everything. It moves the project from documenting a fact to explaining the architecture of grammar.
A strong proposal begins with empirical precision. Before invoking Minimalism, Agree, Case, phases or feature valuation, the researcher must establish exactly what the language does. Which constructions show the pattern? Under what conditions does it occur? Is it categorical or variable? Does it depend on aspect, transitivity, animacy, Case, number, gender, definiteness or argument structure? Are there apparent counterexamples? What happens when one grammatical variable changes while the others remain constant? This is where many proposals fail. They begin with theory before establishing the empirical puzzle. But theory cannot rescue an imprecisely defined phenomenon.
The next step is to formulate the research problem. A research problem is not that “little research has been done.” Nor is it that a phenomenon is “interesting.” A genuine problem exists when an established observation creates an explanatory difficulty.
For example:
Existing accounts successfully describe the distribution of agreement in canonical constructions, but they do not adequately explain why the same agreement relation changes when Case or aspect changes.
Now there is something to investigate. The literature review should then become an investigation of competing explanations, rather than a chronological catalogue of scholars. A weak literature review says what Chomsky, Baker, Preminger or other researchers have written. A strong one asks what each analysis explains, what assumptions it requires, where it succeeds, where it encounters difficulty, and what empirical evidence distinguishes it from its competitors.
This produces the most important section of the proposal: the research gap. The gap is not simply missing literature. It is an unresolved explanatory space. Perhaps one analysis treats agreement as a syntactic dependency but struggles with Case-sensitive variation. Another explains the variation morphologically but leaves the syntactic dependency unexplained. A third captures the data but requires construction-specific stipulations. The research opportunity lies precisely in the tension among these accounts.
At this point, the theoretical framework acquires a genuine purpose. A researcher working within the Minimalist Program should not simply write, “This study adopts a Minimalist framework.” The committee needs to know which components of the theory are necessary to solve the problem.
If the project concerns agreement, the relevant architecture might involve Agree, φ-features, valuation, structural Case, locality and phase-based derivation. If it concerns movement, Internal Merge, feature-driven movement and locality may become central. If it concerns morphology, the researcher may need to examine the relationship between morphosyntactic features and their morphological realization. If the evidence points toward an interaction between syntax and morphology, a theory of the interface may be more appropriate than a purely syntactic account.
The crucial principle is simple:
Do not choose theoretical machinery because it is fashionable. Choose it because the empirical problem requires it.
This also changes how hypotheses should be formulated. In theoretical linguistics, a hypothesis should not merely predict that a particular sentence will be grammatical. It should specify why.
Suppose the hypothesis is that agreement depends on the accessibility of an argument to a syntactic probe. Then the researcher must identify the probe, the potential goals, their structural configuration, their relevant features, and the conditions under which Agree succeeds or fails.
The analysis should make predictions. If Case makes an argument inaccessible to a particular probe, comparable constructions should exhibit corresponding effects. If a proposed asymmetry is purely morphological, syntactic configurations should remain constant while morphological realization changes. If the phenomenon follows from a general Agree mechanism, the analysis should extend beyond the individual construction for which it was initially developed.
This is where the distinction between description and explanation becomes decisive. A descriptive analysis says:- In Construction A, the verb agrees with X; in Construction B, it agrees with Y.
An explanatory analysis asks:- What independently motivated grammatical mechanism makes X accessible in the first derivation and Y accessible in the second?
The second question is what transforms linguistic observation into theory. The same principle applies to morphology. Morphological form should not automatically be treated as a transparent reflection of syntactic structure. A visible suffix may be the realization of a syntactic feature, the result of a morphological rule, the product of syncretism, or the consequence of interactions occurring at the syntax–morphology interface. Consequently, a sophisticated proposal must distinguish at least three questions:
What structure does syntax generate?
What features are established in the derivation?
How are those features morphologically realized?
Confusing these levels can make an apparently elegant analysis theoretically unstable. Methodology in theoretical linguistics should likewise be understood more carefully. A syntax dissertation may not require questionnaires or large-scale statistical modelling, but it still requires a rigorous method. The researcher must specify where the data come from, how constructions are selected, how grammaticality judgments are established, how examples are classified, how competing analyses are evaluated, and how conclusions are generalized.
Corpus data can establish distribution. Native-speaker judgments can establish grammatical contrasts. Published descriptions can provide comparative evidence. Fieldwork can reveal patterns unavailable in existing literature. None of these automatically provides an explanation. The theoretical task begins when the evidence is converted into structural generalizations.
A particularly powerful research design is:
data → classification → generalization → structural representation → feature analysis → derivation → competing analysis → prediction → evaluation.
This procedure prevents the proposal from becoming a collection of attractive examples. The same discipline should govern cross-linguistic research. Comparing English, Urdu and Saraiki, for instance, should not mean writing three independent descriptions. The theoretical question should be why the languages differ if the underlying grammatical architecture is assumed to be broadly shared.
That is where comparative linguistics becomes theoretically valuable.
The goal is not simply:- “English does X, Urdu does Y, and Saraiki does Z.”
The stronger question is:- What grammatical differences are necessary to derive X, Y and Z from a common architecture?
This is the point at which variation becomes evidence about universality. A good PhD proposal must also know what it will not explain. “The syntax of Urdu” is not a dissertation topic; it is a research program. “The interaction of Case and agreement in selected Urdu perfective constructions” may be one. Delimitation is not an admission of weakness. It is evidence that the researcher understands the scale of the problem.
Ultimately, the committee is not asking whether the applicant has read enough books or whether the topic sounds sophisticated. It is asking five deeper questions:
Is there a genuine problem?
Is it theoretically significant?
Can the proposed evidence distinguish among competing explanations?
Does the researcher possess an adequate analytical framework?
If the research succeeds, what will linguistics know that it does not know now?
The final question is the most important. A PhD contribution need not overturn an entire theory. It may be empirical, descriptive, analytical, theoretical, comparative or methodological. But it must be identifiable. “This study will contribute to the field of linguistics” is not a contribution. A claim such as “the analysis demonstrates that the apparent agreement asymmetry follows from the interaction of Case and Agree rather than from construction-specific morphological rules” is a contribution because it changes the explanatory landscape.
This is why the best research proposals are not written by starting with the headings Introduction, Literature Review, Methodology and References and filling them in one by one.
They are built backward from an intellectual problem. First determine what the language is doing. Then determine what existing theories cannot explain. Then determine what evidence would discriminate among the competing explanations. Then formulate the questions. Then select the theoretical machinery. Then design the analysis. Only after that should the conventional sections of the proposal be written.
The proposal is not the research itself. It is the architecture of the argument that makes the research necessary.
A useful final test is to complete this sentence:
This dissertation investigates X because existing analyses cannot adequately explain Y; using Z, it will determine whether A can account for the evidence without B.
If that sentence is precise, the proposal probably has an intellectual center. If it is vague, no amount of additional references will fix it. For the aspiring linguist, this may be the most important transition in doctoral research: learning to stop asking merely what happens in a language and start asking what a linguistic fact forces us to assume about the grammar that produces it.
A Master's project can describe a pattern. A PhD must explain it, and the finest PhD research goes one step further: it shows that a seemingly local fact about a particular language is actually evidence for a much larger question about the architecture of human language itself.
