As with all aspects of language, the sounds of languages and dialects are changing all the time. If you 
look at English spelling, words like knife and cough reflect different earlier pronunciations (say them out loud and think about the spellings!). We can study sound change by comparing written forms of language at different times, e.g. the language of Shakespeare with the language of drama today. It is also possible to study sound change as it is in progress. In order to do this we use knowledge and methods from two areas of linguistics:
Phonetics allows us to describe the production and perception of speech. We can listen to speech and write it down with a special alphabet, the International Phonetic Alphabet (IPA). If you do our two quizzes about pronunciation (“t” or not; disappearing “r”), you can learn how to listen carefully to speech sounds and identify what you hear.
We can also record speech and use speech processing software to measure fine-grained aspects of the speech signal (acoustic phonetic analysis). In the Sounds of the City project, we used acoustic phonetic analysis to identify and track variation and change in Glaswegian pronunciation over the 20th century.
Sociolinguistics is the study of language and society. This branch of linguistics provides the methods to observe sound change in progress. Since recordings began to be made in the early 20th century, there are two methods.
We can compare recordings made at one point in time with those made at another point in time, and see if change has taken place (‘real-time’ change). This can either be from different speakers or from the same speakers over their lifetime. A famous study was recently carried out on the speech of Queen Elizabeth II in her Christmas broadcasts which showed that her speech has subtly changed to become a bit more like than of the general Southern English population.
Another method takes speakers of different ages recorded at one point in time. In general after people have acquired their language around the age of 7-8 years old, most tend not to change it much over their lifetime. This means that listening to the speech of older speakers allows a special ‘window’ on what language sounded like when they were children. By comparing speech in younger and older speakers, we can infer that speech may be changing over time (‘apparent-time’ change).
In the Sounds of the City project, we used both real- and apparent-time methods. We collected recordings from four decades, the 1970s to the 2000s. We also analyzed a very few recordings made in 1916 and 1917 of Glaswegian soldiers in German prisoner of war camps during the First World War, from the amazing Berliner Lautarchiv collection from the British Library. We also compared speakers of different ages from each decade. Our oldest speakers were born in the 1890s and recorded in the 1970s; our youngest were born in the 1990s and recorded in the 2000s. This gives us a span of around 100 years of Glaswegian.
For more links and references about studying sound change, click here.
