Dialects and accents
Dialect refers to all aspects of language typical for some region: grammar, vocabulary and pronunciation. Accent refers only to pronunciation, just one component of dialect. For example, Northern British English (NBE) is one regional dialect of English, spoken north of a line from the Wash to the Severn and roughly as far as the Scottish border. Southern British English (SBE) is another regional dialect, spoken south of the same line, from Cornwall to Norfolk.
Languages typically have local variants spoken within a dialect area, regiolects. SBE includes East Anglian, London (with Cockney), Home Counties and Southwest. Sound changes mean that local accents are constantly changing. For example, 19th century sound changes in industrial towns along the Thames from London to the estuary (Estuary English) had spread to much of Kent and Essex by 1900 and to all the Home Counties (the triangle Sussex-Hampshire-Northants) during the 20th century. These sound changes are still moving westwards, generation by generation. In Northern England, Liverpool shared the same rhotic NBE as most of Lancashire before around 1850, when Scouse appeared, becoming established by about 1900. Then, during the 20th century, Scouse continued spreading along Merseyside into North Wales and towards Manchester.
Further, within any regional or local variety, there might be socially related accents, sociolects, labelled variously by different authors. Jones called them uneducated and educated, an unfortunate choice because uneducated was not defined and educated was related circularly to RP. Wells prefers to call them popular and standard, the endpoints of a continuous sociolect scale. But standard is also a difficult term, implying approval, but with the approver not always being defined. It’s difficult to find a pair of terms where one isn’t pejorative or discriminating. The most neutral I can come up with myself is unmodified and modified, where an unmodified accent X has all the features of X present whereas some have been dropped for the modified accent X. Features are typically dropped in a preferred order along the sociolect scale.
However, the status of national standard accents is a controversial topic among linguists. There are two basic approaches. One is that a national standard accent is somehow the best variety of pronunciation for a language, independent of regional varieties. The other view is that regional features cannot be avoided and that a standard accent is consequently one of the existing regional accents, one of its sociolects. Neither approach solves the problem of who nominates an accent as standard. A typical solution is that an accent is declared standard by its own speakers, as in the case of RP.
Regional differences and sound change
Regional differences in pronunciation usually arise from sound changes spreading gradually and unevenly across the country, so that different areas get out of phase. For example, words like foot, put, soot, strut, but, putt all rhyme in NBE. But not in SBE, since a 16th or 17th century sound change in the south that split those words into two groups there: (a) foot, put, soot, etc. and (b) strut, but, putt. etc.
A useful tool for comparing accents are the lexical sets that resulted from various dialect developments or sound changes (merging, splitting, shifting etc), leading to the regional differences we hear today. For convenience, sets of words like this are referred to by the keywords used by John Wells (1982, Accents of English, C.U.P.). For example, the two particular sets of words mentioned above are known as the FOOT and STRUT sets and the sound change is the FOOT-STRUT split.