A search for “constructive dismissal” can return hundreds of results. A search for whether an employer’s conduct amounted to a repudiatory breach, despite no express dismissal, starts much closer to the legal question. That distinction defines keyword search versus argument search. For Hong Kong practitioners, it can determine whether research produces a list of documents or the authority needed to support a submission.

Keyword search versus argument search: the core difference

Keyword search retrieves documents containing specified words or phrases. It is direct, familiar and often highly effective when the terminology is settled. Searching a statutory section number, a named case, a judge, a distinctive legal test or an uncommon phrase can quickly locate the right source.

Argument search starts with meaning. Rather than asking whether a judgment contains the exact words entered, it asks whether the judgment addresses the legal proposition, factual pattern or line of reasoning behind the query. It is designed for the moments when the language used by the court differs from the language used by the researcher.

The distinction is not between an old method and a new one. Both have a place in serious legal research. The practical question is which method will reduce false leads and surface the most useful authorities for the issue at hand.

When keyword search is the right tool

Keyword search is efficient when there is a reliable anchor. If you know the relevant ordinance, section, case name or technical term, exact matching can provide fast confirmation. It is particularly useful for checking judicial treatment of a specific provision, locating a citation, tracing references to a known authority, or finding a phrase that has acquired a recognised legal meaning.

It also gives the researcher control. Carefully chosen Boolean operators, proximity terms and filters can narrow a substantial database to a manageable result set. For discrete verification tasks, that precision is valuable.

Where keyword search loses time

The limitation is linguistic. Courts do not always describe an issue using the vocabulary in a solicitor’s note, a client’s email or opposing counsel’s pleading. A judgment may analyse the same proposition through different terminology, refer to an earlier formulation of a test, or focus on facts rather than label the legal issue directly.

A researcher looking for “duty to warn” may miss authorities framed around disclosure, informed consent, material risk or the scope of professional advice. Searching for “unreasonable delay” may overlook decisions that assess prejudice, abuse of process, procedural fairness or case management without using that exact phrase.

The usual response is to run more searches. That can work, but it often creates a cycle of synonyms, alternative spellings, exclusions and increasingly long result lists. The time cost is not simply searching. It is reading documents that mention the right words but do not assist the argument.

What argument search changes

Argument search allows the researcher to describe the issue in ordinary legal language. For example: “Can a party rely on a contractual limitation clause where its own fundamental breach prevented performance?” The search system evaluates the legal relationship between the concepts, not only whether every word appears in a result.

This is semantic search applied to legal research. It identifies potentially relevant judgments based on the substance of the query, including related doctrines, judicial reasoning and comparable factual situations. The result is not an answer that replaces legal analysis. It is a more intelligent route to the authorities and passages that require analysis.

For Hong Kong law, jurisdictional focus matters. A broadly trained search tool may identify persuasive material from other common law jurisdictions, but a practitioner first needs to establish the applicable local position, the treatment of binding authority and the relevant statutory framework. Argument search is most useful when it works across dependable Hong Kong case law and legislation, rather than treating legal sources as undifferentiated web content.

Why legal arguments are rarely expressed as keywords

Legal research begins with a client problem, not a database label. The facts may involve a director who diverted an opportunity, a landlord who accepted rent after alleged forfeiture, or an insurer relying on non-disclosure. The legal issue emerges through analysis: fiduciary obligations, waiver, election, materiality, causation or contractual construction.

By the time a researcher selects keywords, they have already made choices about how to characterise the issue. Those choices can narrow the search too early. An argument-based query retains more of the proposition. It can search for the relationship between conduct, legal consequence and governing principle.

This is especially valuable at the start of a matter, when the terminology is uncertain. It is also valuable where the law develops through fact-sensitive application rather than a single statutory definition. Negligence, fiduciary duty, procedural discretion, implied terms and remedies frequently demand that broader contextual approach.

Argument search does have limits. A semantically relevant result may be less useful than a precise keyword result where the question turns on the wording of a defined term, a particular subsection or a procedural rule. It can also return authorities that are conceptually adjacent but distinguishable on a decisive fact. Professional judgement remains the filter.

A more efficient research workflow

The strongest workflow combines both methods rather than treating them as rivals.

Start with the proposition, not the label

State the legal issue as a short question that captures the facts and consequence. Instead of entering “estoppel”, try: “Can a landlord enforce a right of re-entry after representing that late payment would be accepted?” This gives argument search enough context to identify cases dealing with waiver, representation, reliance and contractual rights.

Review the leading results for the governing test, relevant Hong Kong authorities and the passages where the court applied the principle. At this stage, the objective is to map the field, not to collect every possible citation.

Use keywords to verify and deepen the result

Once the search identifies a leading authority, use conventional search techniques to trace it. Search its citation, key judicial formulation, statutory provision and later treatment. This is where keyword search becomes highly efficient: it helps confirm currency, locate closely related decisions and identify whether a specific phrase has been adopted or limited.

The sequence matters. Argument search can reduce the risk of starting with the wrong terminology; keyword search can then sharpen the research around the correct authorities.

Read the decisive passage in context

A relevant case name is not a usable authority. Check the ratio, the facts, the court level, the date and the proposition actually decided. Read the paragraph before and after an apparently favourable extract. A statement that looks broad in isolation may be qualified by the facts, the procedural posture or a distinction from earlier authority.

This is also where tools such as AI-generated case summaries and key passage extraction can save time, provided the underlying judgment remains available for verification. Common Laws.ai supports this workflow by pairing semantic discovery with source-based research across Hong Kong legal materials.

When argument search delivers the greatest value

Argument search is particularly effective when a matter has unfamiliar facts, the client’s language does not match legal terminology, or the issue crosses doctrinal categories. It can expose authorities that a search limited to one label would miss.

Consider an application concerning a delay in commencing proceedings. A keyword-only approach may focus on limitation. An argument search can bring forward decisions dealing with prejudice caused by delay, abuse of process, evidential unfairness and discretionary case management. The final answer may still depend on a particular limitation provision, but the wider authority base improves the analysis.

It is equally useful in legislative research. A query can begin with the practical question and then lead the researcher towards the relevant provision, amendments and judicial interpretation. Where legislation has changed, however, point-in-time checking is essential. A result is only reliable if the version of the provision matches the date relevant to the dispute.

Precision is the standard, not just speed

Faster research is valuable only if it improves the quality of the legal work that follows. Keyword search offers exactness where the researcher knows what to find. Argument search offers breadth of legal meaning where the issue is still being framed. Used together, they reduce blind spots without sacrificing control.

The better question is not whether a search tool found the words entered. It is whether it helped identify the authority, reasoning and statutory context needed to make the next legal judgement with confidence.