Law librarians and legal researchers are perhaps the last people that need telling ‘words are important’– not only do we spend a vast amount of the working week reading and interpreting texts, but the profession often attracts those who enjoy reading for pleasure. Yet even though as a group we are aware in general of the importance of words, it can be easy to forget in the moment the impact of our choice of words, because doing something repeatedly often leads to complacency.
You might receive a research request similar to one from 18 months ago, and without realising it, you could end up repeating the same approach– using the same words and resources– even if it’s not fully relevant to the current request. But during those 18 months political, cultural and social changes could have introduced a number of new terms and phrases that need consideration.
New words are used and recognised all the time– in the most recent update to the Oxford English Dictionary, nearly 600 words, phrases and senses were added, including beating heart, busy bee, secret admirer, ditto, and cold turkey. Footnote 1 Global events and technological advances have also led to spikes in the creation of new words and phrases. The Covid-19 pandemic introduced terms such as social distancing and antivaxxer into mainstream use; a request for research into public health legislation may not have needed to factor in social distancing as a search term pre-Covid, but post-Covid it may be a key search term.
Similarly, the increasing capabilities of artificial intelligence (AI) have created a whole new context in which to use hallucinate. Once just used to refer to sensory delusions in humans, it may now apply to cases of AI systems generating information that appears plausible yet is entirely fabricated or inaccurate.
Just as modern global events have reshaped common language and related research strategies, there have been several times throughout history where the choice to use a particular word has had a profound and wide-reaching effect.
EMS DISPATCH
In 1870, tensions were high between France and Prussia. France was in a precarious position with growing republican opposition to the imperial regime, and recent military endeavours against Austria and Mexico had been underwhelming. Prussia, on the other hand, was emerging as a dominant power in Europe, having recently consolidated its power under the North German Confederation, and it had had several military victories against Austria.Footnote 2
Following the deposition of Queen Isabella II of Spain in the Revolution of 1868, the Spanish Army invited Leopold, Prince Hohenzollern, to take up the throne. Although Prince Hohenzollern renounced his claim to the throne, France feared encirclement by the increasingly powerful Prussia. French Emperor Napoleon III instructed his ambassador to Prussia to demand that King Wilhelm Iof Prussia declare he would never permit a member of the Hohenzollern family to sit on the Spanish throne.Footnote 3
On 9 July 1870 Vincent Benedetti, French ambassador to Prussia, met King Wilhelm in Ems, a resort spa east of Koblenz on the Lahn River. Wilhelm refused this request and after the meeting sent a telegram to his Chancellor, Otto von Bismarck, to report the demands made by Benedetti on behalf of France.Footnote 4
Bismarck made edits to the telegram before it was published in German newspapers to make it more provocative, with his key change being the use of the word ‘adjuant’. In German ‘adjuant’ refers to a staff officer, but in French it refers to an aide de camp (a non-commissioned rank).Footnote 5 When the telegram was published in the French newspapers on 14 July 1870, ‘adjuant’ was seen as Wilhelm insulting the French ambassador and thereby France itself.Footnote 6 In France this resulted in a number of nationalist demonstrations and pro-war publications in the press continued. The Government began calling up reservists and there were several attacks on the Prussian embassy. Though not the only factor, the provocative wording in the Ems Dispatch certainly had an incendiary effect and contributed to the declaration of war on 19 July 1870.
The Franco-Prussian war ultimately lasted for six months and led to the deaths of approximately 180,000 soldiers. It overturned the geo-political traditions of Europe and led directly to the creation of the German Empire, while it inspired a number of military and weaponry advancements.Footnote 7
POTSDAM DECLARATION
In the summer of 1945 after the defeat of Nazi Germany, Allied leaders convened in Potsdam to discuss the administration of Germany and the ongoing fighting against Japan.Footnote 8 The Potsdam Declaration was issued by the United States, Britain and China demanding the complete surrender of Japan lest it face “prompt and utter destruction”. Following the Japanese Government’s discussions of the terms, Prime Minister Kantaro Suzuki stated that the Japanese policy towards the declaration was one of “mokusatsu”.
In a November 1950 article published in the Pacific Historical Review, Jazuo Kawai argued that Japan’s response to the Declaration had been misunderstood because of the translation of mokusatsu – he maintained that it should have been interpreted as “killing it with silence”, but many translators interpreted it as “ignore”.
Until recently it was widely held that Japan’s unintended rejection of the Potsdam Declaration was used by the United States as justification for its atomic attacks against the cities of Hiroshima and Nagasaki, and by the Soviet Union as a reason to invade Manchuria.Footnote 9 Recent academic research into the incident has suggested that mokusatsu is not an ambiguous term and that Prime Minister Suzuki’s response was intended as a flat-out rejection of the Potsdam Declaration.Footnote 10
WORDS IN PRACTICE
The importance of language and its impact is not limited to global political events. In legal research, the nuances of word choice can have a significant effect on research scope and results. Choosing the right word to refer to the person bringing the claim in theory seems like a straightforward decision, but knowing whether to use ‘claimant’, ‘applicant’ or ‘petitioner’ can have a large impact on the volume and jurisdiction of cases returned when conducting a search.
Westlaw case search carried out 18 August 2025
The different applications of each term are often complex, and usually influenced by the area of law and/or the court jurisdiction. Apertinent example would be in insolvency cases, where the type of proceeding may impact which term is used; a person presenting a winding-up or bankruptcy petition is a ‘petitioner’ but a person bringing any other type of insolvency proceeding is an ‘applicant’.Footnote 11 The date of the proceedings may also be a factor, as in family cases prior to 6 April 2022 the person applying for a divorce or annulment would be referred to as a ‘petitioner’, but after 6 April 2022 would be known as an ‘applicant’.Footnote 12
CHOOSING THE RIGHT WORDS
So if our word choices will have such a large impact on the outcome of our searches, how do we make sure we are choosing the right words?
CONTEXT
Contextualising a research request provides valuable information that helps us make informed word choices. Sometimes a requestor may ask something very broad such as “cases related to financial sanctions”, or in their haste to send you the request between calls, they may quickly type out an email mentioning “the applicant in a bankruptcy appeal” or referring to the “petitioner in a new divorce matter”. The phrasing used by requestors may not always be the exact terminology you need to conduct your research, and should not be taken as definitive.
Establishing further context in these two examples can be done in three ways: having a baseline general knowledge of the legal system; asking focused follow up questions to the requestor; and reading around the specific subject before launching into the research process.
General knowledge of the legal system can be taught through law degrees, industry training courses and internal new starter programmes, but is perhaps most valuably obtained through day-to-day experiences of working in law, and having a buddy or mentor that can pass on the nuances of the legal system as and when questions arise. Through previous experience of insolvency cases, you might recognise that ‘applicant’ would not be the right term, or that bankruptcy appeals are heard by the Chancery Division.
Asking focused and tailored follow up questions to the requestor can narrow down vague requests. If we return to our bankruptcy example, knowing that the requestor is looking for appeals would help us determine that ‘petitioner’ is the correct way to refer to the person bringing the claim. Our searches could be narrowed down even further by knowing if the requestor has a timeframe in mind, is interested in sanctions relating to a specific country, or is looking for cases against individuals and not companies. It would also save valuable time by checking what (if any) research the requestor carried out before approaching you, to avoid duplication of work and wasting time.
A list of basic questions could be provided to a new starter, but asking good follow-up questions is refined through on-the-job experience. Questions you may have asked six months ago when new to the job may no longer need asking, as you know some of the context already.
Taking a short amount of time to get a basic understanding of the topic will help focus your research and help pinpoint key words you should be using. Introductory practice notes and guides offer quick and easily digestible overviews of complex legal topics– both Practical Law and Lexis+ have introduction or new starter sections that feature glossaries, flowcharts of complex procedures, step-by-step guides and links to other useful resources. Many other platforms also have content that can be extremely useful to those new to a subject, such as FromCounsel’s ‘skimmed’ Q&A or Lexology’s Panoramic Q&A.
KEYWORD MAPPING
When a request comes in, the gut reaction is often to immediately start working on it, especially if it has been flagged as urgent. However, taking just five minutes to map relevant words and phrases at the start of a research task can save time and energy later in the process.
Perhaps the urgent request has been for cases where businesses have objected to local infrastructure frameworks. Through contextual questioning you may narrow down the search parameters, having established that the requestor is looking for examples within the last year, in England only, and is interested in cases heard in the Kings Bench Division (Technology & Construction) Court only.
By taking the time to think of other ways to refer to ‘framework’, it would ensure you do not miss a case where instead of a framework the judgment refers to a scheme instead. Then armed with a list of potential terms, you could systematically work through the list, leaving your base contextual search criteria in place and merely changing the key word each time.
As the following example shows, there are varying numbers of cases returned just by changing the key word searched. It’s likely that there is a large amount of overlap between the cases returned, but through further investigation of the cases those that are most relevant would be identified.
Westlaw case search carried out 18 August 2025
AI
It’s also possible to use artificial intelligence (AI) to support our choice of words during research, if done with care and caution. Numerous recent news stories have highlighted the pitfalls of relying too heavily on AI when conducting legal research– examples of poorly drafted contracts and pleadings citing fictitious cases have raised serious concerns over how the legal industry uses AI.
However, if researchers are deliberate in their phrasing of prompts, follow their organisational AI policies and do not blindly trust its output, it can be a useful aid alongside your existing practices.
“Explain this to me as though speaking to a 12-year-old …”
AI can be used when we’re trying to establish the context of a request. By asking the chosen AI model to explain a subject to someone with a basic understanding, or to a 12-year-old, it will break down a complicated topic into its most essential components. When asked to explain the trust law concept of gifts with reservation of benefit as though talking to a 12-year-old, Microsoft Copilot creates an analogy of a grandparent giving a grandchild a house to explain that a person gifting something, but still using it or benefiting from it, means that the law treats it as if they still own it for tax purposes. The response removes complex jargon and uses basic vocabulary, shorter sentences, and familiar everyday concepts to quickly explain the topic, while linking to the source material for further reading and fact checking.
Using this method to get a basic understanding of a topic is especially helpful prior to reading practice notes or book chapters, which are often written under the assumption that the reader already has a basic understanding of the concept.
“Help me think of other ways to phrase …”
AI can also be used to support key word mapping. If we look at a request for documents related to autonomous vehicles, if the topic is completely new to the researcher not all relevant and interchangeable words might be obvious. AI can generate lists of synonyms, regional variations or domain-specific terminology. It can also help challenge assumptions, ensuring broader and more inclusive search results.
Including contextual information relevant to the request helps focus the AI’s output. An example may be asking for alternative phrases for ‘autonomous vehicles’ that are used in Europe. The researcher may have already read around the subject and compiled a list of related words or other phrases often interchangeable for ‘autonomous vehicle’. AI could then be used to generate a list of regional variations and their literal translations, to account for variations in the style of translation– the Italian phrase ‘auto a guida autonoma’ would often be translated as ‘self-driving car’ but could be translated in the very literal sense as ‘car with autonomous driving’, which could be missed when searching translated texts if the phrase ‘autonomous driving’ is not a search term (Figure 1).
Figure 1: Microsoft Copilot, 14 July 2025
“Review and improve the following Boolean query …”
Many researchers will have experienced the fatigue that can occur after working on a complex Boolean query– the knowledge that somewhere within the query there is a missing quotation mark or a wrong connector, but you are no longer able to see the error. Using a generative AI model to identify these mistakes with a simple prompt of “You are a researcher experienced in using Boolean logic. Review and improve the following Boolean query” can save valuable time by refining your grouping, ensuring logical clarity and removing redundant or ambiguous terms.
By adding a few extra words into the prompt, the AI can also become a learning tool– “you are a researcher experienced in using Boolean logic. Review and improve the following Boolean query. Outline the changes and improvements you have made.” The AI model can provide step by step guidance on the changes it has made to your initial query, referencing each change it has made so that you are easily able to follow its revisions.
Figure 2 is an example of inputting the above prompt into Microsoft Copilot.
Figure 2: Microsoft Copilot, 18 August 2025
CONCLUSION
The impact of language in legal research is both practical and profound. Historical episodes like the Ems Dispatch and the Potsdam Declaration provide a reminder of how nuanced shifts in wording can influence the course of global events. In the legal domain, a single variation in terminology– whether ‘claimant’, ‘applicant’, or ‘petitioner’– can dramatically alter the scope and relevance of search results. Effective legal research depends on deliberate word choice informed by context, supported by thoughtful consideration of alternative phrasing, and enhanced through careful use of AI tools. Legal information professionals must remain alert to linguistic precision– in doing so, they not only improve the accuracy and depth of their research but also strengthen the quality of legal practice itself. Words matter– in every request, every search, and every outcome.
