In varieties of Southern States English and African-American English, the vowel sounds 
and 
merge when they occur before nasal consonants. As a result, words like pin (SAE 
) and pen (SAE 
) both have the pronunciation in Southern States English.
It has been theorized that the merger of and before nasals has its sources in British English (Krapp 1925; Kurath and McDavid 1961; Morgan 1969). Morgan (1969) states that "the substitution of the vowel 
for 
as in pin for pen" is derived from the British Isles and is common in the speech of North Carolina. Kurath and McDavid (1961) state the following:
"In English folk speech again as the of pen in the central counties...[and] the of pin in the eastern counties...The variants existed in Middle English, survive in English fold speech of today, and had a social rating in London English of the eighteenth century not unlike that in present-day American English."
Krapp (1925) also maintains that this phenomenon (which he finds in southwestern Virginia) is a relic of seventeenth-century colonial English.
Brown (1991) discusses the phenomenon of merger in Tennessee, with a particular focus on the spread of the phenomenon throughout the state since before the Civil War. She uses three primary sources for her data in this study. One source is a collection of questionnaires collected by a Tennessee archivist Gus Dwyer between 1918 and 1922. These questionnaires were collected from Tennessee veterans of the Civil War, and thus represent informants that were born between the years of 1820 and 1850. Brown uses the spellings of words such as friendly as frindly to indicate instances of an informant that has merger. A second source for this study comes Tennessee surveys from LAGS, or the Linguistic Atlas of the Gulf States, the data in which was collected more recently and represents the speech of informants that were born between the years of 1872 and 1960. The third source was designed to fill in the gap between the questionnaires and the LAGS data. This data was from the North Carolina surveys of LAMSAS, or the Linguistic Atlas of the Middle South Atlantic States. Since the speech of North Carolina is very similar to the speech of Tennesee, and since there is data that is collected from informants born between the years of 1840 and 1900, Brown proposed that the insights into the evolution of merger during that time period can be extrapolated for Tennessee during the same time period.
From these three sources, Brown has data pertaining to merger from informants whose birthdates range from 1820 to 1960. The study of these data have interesting results that may shed some light on how this phenomenon spread throughout Tennessee. Brown's study of the Civil War veterans questionnaires revealed the following statistics:
| Year of Birth | Merger | No Merger | |
| Tennessee Veterans | (1820-1839) | 14 (14.9%) | 80 (85.1%) |
| Tennessee Veterans | (1840-1851) | 26 (11.7%) | 197 (88.3%) |
| Veterans Totals | 40 (12.6%) | 277 (87.4%) |
As the above table illustrates, for informants born between the years of 1820 and 1851, there is a higher incidence of no merger in words such as friendly. Brown takes this as an indication that 87.4% of these informants do not have the variant in such words. The following map shows the locations of the informants for this study:
 |
|
 |
 |
As this graphic illustrates, the data collected regarding Civil War veterans seems to indicate that there was variation across most parts of Tennessee, with a few areas of merger.
The LAMSAS data revealed the following statistics for informants in North Carolina:
| Year of Birth | Merger | No Merger | |
| North Carolina LAMSAS | (1840-1855) | 2 (12.5%) | 14 (87.5%) |
| North Carolina LAMSAS | (1856-1871) | 9 (13.8%) | 56 (86.2%) |
| North Carolina LAMSAS | (1872-1900) | 22 (31.0%) | 49 (69.0%) |
| LAMSAS Totals | 33 (21.7%) | 119 (78.3%) |
This table seems to indicate that there is an increase in merger among speakers born between 1872 and 1900 (compare 1840-1855 12.5%/87.5% to 1872-1900 31.0%/69.0%).
The third source, the LAGS data, revealed the following statistics for informants in Tennessee born between the years of 1872 and 1960:
| Year of Birth | Merger | No Merger | |
| Tennessee LAGS | (1872-1900) | 30 (46.2%) | 35 (53.8%) |
| Tennessee LAGS | (1901-1930) | 37 (61.7%) | 23 (38.3%) |
| Tennessee LAGS | (1931-1960) | 31 (91.2%) | 3 (8.8%) |
| LAGS Totals | 98 (61.6%) | 61 (38.4%) |
This table again seems to indicate that there is an increase of merger between the years 1872 and 1960, where merger has increased from 46.2% to 91.2%.
The following graphic is a map illustrating the results of the LAGS survey of the words ten, twenty, and pen:
 |
|
|
|
The graphic above shows that the LAGS survey, which is more recent, finds more instances of merger across Tennessee. Also, Brown points out that the areas of merger are centered around waterways, roadways, and urban centers such as Nashville and Memphis. Consequently, the incidences of non-merger are relegated to outlying areas.
Based on the regional data, Brown proposes that the increase in incidence of merger among speakers may be due to the increase in urbanization in post-Civil War Tennessee. The proposal is that before the Civil War, there were a few incidences of merger among English and Irish settlers of little education. Among other English settlers, the non-merger form was predominant. However, after the Civil War, many rural speakers moved to urban centers to find work in the new manufacturing industries. At that point, the merger phenomenon may have been associated with the working class, but then lost its stigma and became predominant.
The following samples are from a white male speaker over 60 from Eastern Texas. The samples include words containing the phonological contexts in which merger is likely to occur.
...attempt to collect..."
© 2001 The Language Samples Project