Population, mutation, linguistics and the number of surnames

	Topics (Brief Coverage)
	Introduction
	Pronunciation
	Education of the clergy
	Misinterpretation of Regional Dialects
	Spelling
	Recent Developments
	Further Research

If you came to this page directly, then please access
Modern British Surnames

The plethora of surnames in the early Middle Ages

"Although the national population was far lower in the middle ages than it is today, England had a much wider range of surnames at the period of their formation than it had in subsequent centuries"
(Source -David Hey: Family Names and Family History)

Many historians currently believe that the population of England reached 5 or 6 million by the year 1300. However, the effect of the Black Death of 1348-50 and later pestilences meant that it had fallen back to the region of 2.2 to 3 million by 1380. This downward trend continued for another 100 years until the population bottomed out at about 2 million. Recovery was slow, and appreciable increase did not show until the 1530's and 1540s. (David Hey: ibid)

The Black Death Maximum estimates % Mortality             Inadequate data 30-34 35-39 40-44 45-49 50 -

How did the ravages of the Black Death affect the stock of surnames? The general population levels may have gradually recovered, but surnames were more vulnerable to erasure. How many were wiped out completely? Which categorary of surnames was effected the most? What was the effect on individual counties?
Impossible questions to answer, though intriguing ones. I do wonder if a trace of the Black Death can be discerned in the current surname density levels of individual counties (Comparing Derbyshire with Staffordshire, for instance)- though one would need to know the surname compostions of each county prior to the Black Death. Again, highly unlikely to unearth. Just day-dreaming....

Population drop was not the only reason for the surname pool to decrease. Sturges and Hackett modelled the rate of surname extinction over 23 generations. They chose a start date of 1350 to enable them to neglect the effects of the plague. Even without accounting for pestilence, 30% of male lines would fail in the first generation according to their model.

The drop in the number of surnames due to these two reasons must have been phenomenal.
The following might be a reflection of this. A recent study of the surnames of Nidderdale has shown that there were no outstanding common names among the taxpayers of 1379 . However, localised established surnames had certainly achieved dominance by the time of the Lay Subsidy of 1546, for they reached a peak after the Dissolution. This peak continued in Nidderdale for as long as conditions were favourable, but received a setback after the Restoration, when family size declined and the national population growth fell to zero. The dominance of certain names came to an end in the mid eighteenth century. (Turner: Distribution and Persistence of Surnames in a Yorkshire Dale, 1500-1750).

Earlier, Richard McKinley in The Surnames of Lancashire had drawn attention to the fact that although some names had ramified considerably by the 16th and 17th centuries, no name was preponderant in the Lancashire Poll Tax returns of 1377-81.

The number of recorded surnames -in England- did increase post 1540, but mainly due to surname corruption and mutation

Reaney wrote:-

"The modern form of very many of our surnames is due to the spelling of some sixteenth or seventeenth century parson or clerk, or even to one of a later date. It is not a matter of illiteracy in our sense of the word. These parsons who kept the parish registers were men of some education. Their ability to read cannot be questioned, but they had no guide to the spelling of names. It was the printing-press which gradually established a recognised system of spelling. That of Tudor and Stuart England was very different from ours, and the spelling of many of our words is not earlier than Dr Johnson's Dictionary. But there was no recognised spelling for names. A great part of the population was illiterate. Their names were only written at birth or marriage or death, or if they happened to come within the clutches of the law. Then they gave their names orally and the clerk put them into writing as best he could. He wrote them down phonetically, using his own system of spelling, sometimes spelling the same name in different ways at different times"

For example, here is how the surname Partridge was recorded in two sources for Suffolk

One name generating 11 mutations in just 2 sources. Dugdale is said to have found 130 mutations for the name Mainwaring.
George Redmonds has commented "..the number of variant spellings of any given name throughout its history, is infinitely greater than the surviving variants. Moreover, this tendency towards uniformity now is probably greater than the move towards individuality"

There must have been numerous factors in the creation of new surname forms . Here are a few possibilities:-

• Pronunciation changes over time
• The cultural background of the clerk
• Misinterpretation of regional dialects
• Lack of standardised spelling and/or changes in spelling conventions
• Name change through peer pressure

Pronunuciation

The pronunciation of surnames at the time of their creation is entirely different to that of today. Indeed, pronunciation changed markedly between the time of Chaucer and that of Shakespeare, but subsequently only in small ways.

Overall, the biggest changes were in the vowel-system; and the main changes are referred to as the Great Vowel Shift.

The situation was slightly more complicated than this, and anyone wanting to take this further should consult the standard texts.

Did the GVS affect surnames? Certainly in how they were pronounced; but not necessarily in how they were spelt. The Present-Day English spelling reflects Middle English pronunciation : the pronunciation of English before the GVS.

The surname 'Dance' was often recorded as 'Daunce' or 'Dawnce' prior to 1600. It seems to have been pronounced as if to rhyme with haunts. Words like dance, aunt and chance have origins in borrowings into Middle English from French. "They entered English with an au spelling which, according to Dobson (1968 II:786), corresponded to a dipthong that was later monophthongised into a short a in some areas and a long a in other areas". After 1600, the au spelling falls away, as if to reflect that the vowel has been shortened, even in the south where the name is predominantly found. This echoes the fact that there was a sporadic shortening of vowels in words of one syllable (especially if they ended in a single consonant). Often short and long variants would be in circulation simultaneously.
Anthea Fraser Gupta 'Baths and Becks' English Today 81(Jan 2005) quoting Dobson

I have concentrated on the GVS because I think it is instructive to learn how our surnames were actually pronounced at different times in the past. However, I would like to emphasise that it is not clear what effect the GVS had (if any) on the recording of surnames. Surname changes were occurring well before the GVS, as the following examples recorded by Reaney reveal :-

Further Reading:-

Charles Barber The English Language : a Historical Introduction (Cambridge University Press, 1993)
H C Wyld A History of Modern Colloquial English (T Fisher Unwin Ltd, 1920)
P H Reaney The Origin of English Surnames (Routledge and Kegan Paul, 1967)

Education of the clergy

Until the end of the 16th century, the standard of education of the average clergyman was low. Inadequate remuneration failed to attract an educated clergy. But by the 1630's the majority of parishes had an resident, graduate clergy. Is this change in education/cultural background evidenced in the recording of surnames in parish registers?

Moreover, the new graduate clergy would have been heavily influenced by Latin. In the 16th century, Latin influence caused words to be remodelled according to their real or supposed Latin etymology. Thus b was introduced into debt (originally dette), p into receipt, and c into indict. Not only spellings, but also pronunciation of individual words altered under Latin influence e.g. aventure (= at a venture) became adventure; verdict from verdit; perfect from parfit.

One is left wondering whether the same happened with surnames? The clergy are usually seen as the honest interpreters of a surname, but could they have been consciously/unconsciously transmuting names?

There is an intriguing reference in Pounds A History of the English Parish p166

A priest at St Ewe (Cornw) moved to another living because he could not understand
the Cornish language, and there are other instances of inability to cope with the local patois

This example came from the 14th century, and a study of the geographical mobility of priests (through locative surnames) at that time reveals that the majority remained in the same diocese, and not infrequently only moved a short distance from their place of birth. However, a significant minority (25%?) had a surprising degree of mobility - e.g. from the deanery of the East Riding of Yorkshire to Cornwall, Devon, Dorset, Wiltshire, Hampshire, Surrey, Kent ,as examples of the furthest interchanges.

Are there any studies of the geographical mobility of parochial clergy in later centuries, and the difficulties they had in understanding?
George Redmonds also gives examples in Surnames and Genealogy of the local clergy humorously altering surnames; deliberately misspelling as a kind of private joke. Wonderful.....

Another source for the creation of temporary variants could be due to hearing loss of an individual cleric. We tend to naturally assume that those recording events had perfect hearing. Not necessarily so. Although it could be argued that this particular defect was corrected through lip-reading. If it is a possibility, then research has been conducted into those consonants that become progressively confused due to hearing impairment. For example, the consonants d,g,b or v,th,z are indistinguishable to those with even a slight hearing loss.

Misinterpretation of regional dialects

Migration is often cited as the single most important cause of surname corruption.

"My MOWBRAY name has variants such as:
MOWBERRY...this was from a Leicester family that went to Northants in the early 1700s. The parish clerk, never having had a MOWBRAY family in his region before, heard the name and wrote what he thought the man had said. The offspring migrated to Lincolnshire, which is prime MOWBRAY territory, and some maintained the variant whereas others reverted to the original. In medieval times, an offspring of the baronial line went to Scotland and founded a dynasty. They adopted the spelling MOUBRAY which has persisted. One can hear the Scottish accent in this variant. In Yorkshire the variant MAWBRAY crops up, this time reflecting the Yorkshire drawl." (Source: David Mowbray email to the Guild Forum)

"In fact this change in the way the name is written down is quite easy to map, and SW of a line drawn between Kidderminster and Redditch it is ROWBERRY and NE of that line it is RUBERY. This is most neatly demonstrated by a 19th Century Ag Lab family who moved backwards and forwards over that line. In the one area their children's births were registered as ROWBERRYs but in the other as RUBERYs." (Source: Polly Rubery email to the Guild Forum)

In the above examples it clearly is. However, how typical are its migration patterns?

Mobility - Mediaeval

There have been several useful studies of mediaeval migration, and all indicate a high level of mobility.

. Twenty Five percent were moving over 20 miles, and a few up to 150 miles. It would be useful to correlate this with points at which dialect certainly changes. The distance for dialect changes is greater for lowland than upland regions: but I hazard a guess at 70 miles (lowland) and 40 miles (upland). Someone more knowledgeable, enlighten me please.
And a higher percentage of the the names that moved would be more distinctive than later, and therefore a higher percentage subject to change.

Post-medieval mobility

David Hey writes that most migration was within a radius of 25 miles; often centring on the local market-town : subsistence migration declining rapidly with the passing of the 1662 Settlement Acts.

For further information, read Ian Whyte 'Migration and Society in Britain, 1550-1830'

Spelling

It is commonly accepted that for the printed word, the spelling system became standardised by the end of the seventeenth century. Powerful influences had been the introduction of the printing press (with spelling conventions set by the printers) and the publication of dictionaries.
This standardisation was not echoed in the recording of surnames. Famously, Shakespeare had 5 other variants of his name, in his lifetime. Indeed, standardisation did not arrive until the advent of mass literacy in the twentieth century. In my initial analysis of the 1900 Portsmouth Burgess Rolls, I am struck by the lack of name standardisation. And this in a section of the populace that one would expect to have a higher incidence of literacy.

Common spelling changes

Spelling changes in names

In surname studies, there is a complicated relationship between pronunciation, orthography and -presumably- the influence of the standardisation conventions of printed text.

Modern Spelling

As mentioned, Surnames have resisted -or have been placed outside- the drive for spelling standardisation. (In contrast to forenames, where variance is allowed, but is very constricted e.g. Catherine, Katherine, Kathryn)
This is perhaps due to the perceived need for surnames to be differentiated from common nouns, in a way other than capitalisation. This process seems to apply particularly to succinct names, and is achieved through a process of padding. This padding is found in 3 forms:-

1. The addition of a superfluous letter 'e'
2. Consonant doubling
3. The substitution of the letter 'y' for 'i'

• In some cases, this trend has become fixed. The surname 'Taylor' is a notable example : this form has superceded the original 'Tailor'.
• In others, both the original noun-form and a spelling variation exist side by side e.g. 'Daw' and 'Dawe'.
• Intriguingly, some common noun surnames have resisted modification e.g. notably, the surnames 'Black' and 'Flint'
  Colours as surnames seem particularly resilient to amendment e.g. Black, Grey, Brown, White. With the latter, it might be argued that 'Whyte' is an exception- but this is predominantly found in Scotland, where it actually outnumbers 'White'. Other influences are presumably in play. Blue, incidentally, is predominant in the west coast of Scotland (Bute and Argyll).

As a counter argument, it must be noted that many of the variable padded surnames are of a lower frequency than the main form.
This 'Crosse' is far rarer than 'Cross'; and similarly for Webbe, Locke, Steele, Brooke etc. The situation is reversed with consonant-doubled surnames, which have become the norm- especially the plural form e.g. Gib-Gibb-Gibbs

1. Superfluous <e>

The addition of the letter <e> can be discerned in 4 types

1. After a short vowel and a consonant-letter cluster

Examples:-

2. After <oo> and a consonant (usually /k/)

Examples: Brook(e), Cook(e), Hook(e)

3. After a long vowel (spelt with more than 1 letter)

Examples:-

4. Vowel spellings with <r>

Examples:-

2. Consonant doubling

This is particularly seen with the following (end) consonants
p b t d g m n and r

Examples:

3. Letter 'Y' Substitution

Examples would be Taylor, Smythe, Pye, Wylde.
In some cases, these forms have predominated because they are seen as having archaic value, in the same sense that the form 'Ye Olde Tea-Shoppe' is used. The substitution adds to the written sense that this type of name is of historic longevity.

Combinations
All the above examples include instances of names that combine 2 types, e.g. consonant doubling and the superfluous <e> e.g. Lily vs Lilley, Crab vs Crabbe

Idiosyncratic spellings
Spellings in which there are no correspondences between spelling and pronunciation
Examples:
Batchelor, Beauchamp, Cholmondeley, Featherstonehaugh, Marjoribanks, St John

Sources: E.Carney 'A survey of English spelling / V Cook 'The English writing system'

Recent developments

The best in-depth analysis of how/why surnames mutate has been George Redmonds Surnames and Genealogy : A New Approach. Dr Redmonds shows -through his study of the use of aliases in Yorkshire parish records- the depth of uncertainty that parish clerks had in understanding and recording local names. Previously, the extent of name changes had been regarded as a minor phenomenon ; Dr Redmonds demonstrates that it was a process of major importance in enlarging the stock of surnames in Yorkshire.

"George Redmonds' Yorkshire examples show that, although some apparent variations in names were simply misspellings by clerks (or in some cases humorous adaptations by clergymen), many surnames changed permanently. Some alterations were minor ones, involving the dropping of aspirates or of consonants such as a final 'd' or 't', but others were more basic because of the the different speech customs of the district into which a family or individual moved" (Source: David Hey 'Family Names and Family History' p84)

Surnames were being changed through peer pressure, into new forms that bore no resemblance to the original name.

For example Cowgill being tranformed into Coldhill. Such transformations could be temporary, and revert on future migration; or could become permanent.

Andrew Todd in his 'Shadows of Ancestors' has summarised some of Dr Redmonds' earlier findings; and identifies the following processes:-

Names with the following structure are liable to be affected:-

Qu/W

S Carlsson in Studies on Middle English Local By-Names in East Anglia (Lund, 1989) p 161 averred that the voicing of W and Wh by Q was a distinctively northern phenomenon, particularly characteristic of Lancashire.. (Although it does occur sporadically outside of 'the north', e.g. Robert Peperquit in a 13th century Sussex charter)

K/C substitution (velar and palatal plosives)

The consonantal substitution of k for c in northern zones is a familiar phenomenon of late Middle English. This was not just a spelling change, but a voiced one as well, in some surnames and by-names
e.g. Godrik (south) vs Godrich (north)

K/C insertion

e.g.Sklater < Slater
Skaithlock < Scathelok

If you have a more than passing interest in surname variance, then you must read Todd and above all George Redmonds. I have included an Appendix which lists changes in much finer detail.

For a discussion of what might constitute a variant, please read Peter Christian's various papers, which are held on the Thesaurus of British surnames website.

How can a true variant be determined?
A difficult question. However, it is possible to say that B cannot be a variant of A, because it would then violate the rules of the sounds of English. [Peter demonstrated this in his paper 'What surname distribution cannot teach us']

Sources:-

• George Redmonds Surnames and Genealogy New England Historic Genealogical Society, 1997
• George Redmonds Migration and the Linguistic Development of Surnames (Family History Vol 11 No 77/78 Aug 1980)
• Andrew Todd Shadows of Ancestors Allen and Todd 1996

Further Research.

Variant or Deviant?

The Guild of One-Name Studies defines a variant as

"a name spelling which varies from the primary name spelling (or another variant spelling) used by that person’s ancestors and which is:
A name spelling that the person was known to have used, through signature evidence on wills, marriage bonds, marriage certificates or other documents originating from the individual concerned
Or A name used by officials on a consistent and persistent basis over a number of years"

The Guild of One-Name Studies defines a deviant as

“any other spelling recorded, including cases where the spelling occurs in official records, but only randomly and inconsistently.
Deviants will also include spellings derived from enumeration, transcription and indexing errors, both contemporary and modern.”

Do names that are subject to drastic variance share some or all of the following aspects?:-

• Local, distinctive (possibly locative= placename surnames)
• Ramify suddenly
• Migrate over dialect boundaries
• Polysyllabic = each syllable being meaningless

There is much work to be done in :-

• Measuring the distance from origin of new surname mutations
• Plumbing the linguistic possibilities of parish registers with regard to the duration of each incumbency

Variations in Reaney

Reaney¹ gives no indication of how he chose to group variants. How did he decide on a head-form? -on etymological grounds?- the most occurrences? the earliest form? the latest form?
With these uncertainties, a peremptory analysis of variants using Reaney is bound to be somewhat suspect.
I spent a week at a hospital bedside, and to pass away the time, I decided to count how many of the variants in Reaney (3rd ed) differed in the lead consonant from the head-form. Most variants are created through a change in a medial vowel sound, or through a doubling of consonants, or through the addition of a suffix. The initial consonant is deemed to carry most 'information' about a name, and hence if it changes, that should be significant.
In the following table, column 1 is the letter of the head-form, and the subsequent columns indicate the count of variants that do not match that initial letter.

Provisos

1. The analysis should have been on phonemes rather than consonants e.g. P's will cover both a hard 'P' and 'F' sounds as in Paul and Phillips. There is grapheme-to-phoneme guidance which helps². Even so, names often do not adhere to these rules, and a personal judgment would still have to be made as to pronunciation
2. Correspondences with D, L, M or O can be explained in some cases with the dropped suffixes -De, Le, Mc and O'

Analysis

• Names that are most likely to 'mutate' are those commencing with A, then G then F..E,C,H,O
• L, R, Tand Yare the initial consonants least likely to be altered. In the case of L, all the changed initial consonants are due to the dropping of the prefix De. The resistance of the letter 'L' may be due to the fact that phonetically, it belongs in a class of its own, as a lateral (where the sound is uniquely made along the sides of the mouth), and therefore less likely to be confusable with other consonants
• If consonants and glides are considered together, then they are not grouped at the top ranks -as might be expected- but throughout the ranking
• Voiced and unvoiced consonants e.g.T and D, F and V, P and D do not appear in groups.
• Variants do not seem to be evenly distributed between similar confusable consonants. There appears to be a movement from one to the other
  e.g. Under F, Reaney groups 58 names beginning with V, but for V, only 12 F'S are listed
• The most common changed-into consonant is 'H' - primarily from A and E
• Is this a result of Reaney's 'system' - or does it indicate something more significant about Variant inial letters?

Appendix: Yorkshire Changes noted by George Redmonds

Appendix 2

Linguistics and the corpus of modern surnames

Only certain combinations of phonemes are possible; and the study of the possible combinations is called phonotactics.

Syllables

1) Number of syllables

Surnames are nouns (in fact a subset of proper nouns). How does their average number of syllables compare with those of ordinary nouns?
A simple analysis of the leading 300 surnames reveals

Compare this with an analysis of the 19,334 nouns in the Longman's dictionary of contemporary English

English surnames (or at least those in the top 300) are predominantly bi-syllabic. This categorary -plus the mono-syllabics- account for 91% of these surnames. Whereas these two categories in the nouns sample accounted for a lesser figure- 60% . Moreover, the surnames sample had no occurrences of surnames with 4 or more syllables; even those with 3 were a definite minority. So surnames can be differentiated from nouns by the pattern of their syllable length.

The English form of identity -it seems- requires succinctness, but not necessarily brevity

As a double-check, I counted syllables from samples of 100 names each from lower-frequency surnames

Bi-syllabic names still predominate at these lower frequencies

The same proportions can be discerned in an analysis of the average syllable length in forenames:-

In the case of monosyllabic surnames, there almost seems to be a 4 letter rule for English surnames.
However, there are exceptions : Cam, Esk, Ham, Kyd, Lot

Source: Anne Cutler, James McQueen, Ken Robinson
Elizabeth and John : sound patterns of men's and women's names Journal of Linguistics 26 (1990) p477

Strong and weak syllables

The English language "consists predominantly of words beginning with strong syllables, and the most common word pattern is in English is a bi-syllabic with initial stress, such as common, pattern, English" ¹

In their study of 190,000 words of spontaneous spoken British conversation, Cutler and Carter (1987) ³ found that 90% of the lexical words (i.e. nouns,adjectives, verbs and adverbs) began with strong syllables. Whereas, considering the total number of weak syllables, 69% were initial syllables of function words, 26% were not word-initial, and 5% were initial syllables of lexical words. [Examples of function words would be articles (e.g., an, the), prepositions (to, from), conjunctions (and, but), determiners (all, those), complementizers (if, that), auxiliary verbs, and personal pronouns.]

Thus the patterns of word stress have resulted with the great majority of English lexical words, like nouns, placing stress on the first syllable. It is the case that a large number of lexical words have only one syllable, but still that one syllable is likely to have stress.

^{For Example}
the predominant stress pattern for bisyllabic words in English is strong\ weak (e.g. m ommy) : weak\strong words
(e.g. all ow) are comparatively rare

Moreover:-

"English words generally follow a pattern of alternating strong and weak syllables. In general, strong syllables are pronounced with a slightly higher pitch and for a slightly longer duration than weak syllables. Strong syllables are long. Weak syllables are usually short. " ²

^{More Examples}

(^. a dot = short, - a dash = long)

²These great examples (and quote) borrowed from Language Center (Hong Kong University of Science and Technology)- I hope to replace them as I become more competent

Surnames as poetry?

• The strong and weak syllables of English result in a rhythm that is similar to the rhythm of poetry. The strong (stressed) syllables are like the beat in poetry. So does the typical English surname have a trochaic foot , in unison with a lexical word???

I do not know - and have not the knowledge to successfully to survey and analyse a corpus of surnames (as spoken). Might there be a difference with nouns?

The other way in which a name might be considered poetic is in the possible euphony through forename choice

• in the choice made for forename-surname combination (e.g. I have a mono-syllablic weak-stressed surname, preceded by a bi-syllabic strong-weak stressed forename). Conversely, for the more normal bi-syllabic surname (with inital strong stress), examples abound of parents choosing a monosyllabic forename.
  Compare Clint Bonney with Ian Bonney.

• The latter has more euphony (i.e. is more harmonious to the ear), because it maintains the common rhythm of alternating strong/weak syllables. One school of linguists argue that the strong syllable is used by the hearer as a first-pass attempt to aurally decide where the division between sounds should be made.¹ This is called segmentation. Parents in their choice may be unconsciously influenced: that Ian Bonney will more easily segmented into forename-surname: whilst the Clint example could fail to be segmented, and consequently (and constantly) be interpreted as the surname -Clintbonney
  This is more clearly exemplified by a name commencing with a weak syllable- such as Patrick.
  If this name(as a surname) is preceded by an initially-strong mono-syllable forename (e.g. Kirk), there is difficulty in segmenting this combination, and to most listeners it would be interpreted as the surname Kirkpatrick, despite an emphasised pause after the first syllable.

• This might have had more significance in the past age of illiteracy and scribal/incumbent recording of names

• As regards surnames alone- what might be revealing is whether the distinctive surnames of a region show a tendency to one of the above rhythmic patterns, in direct comparison to another region. (Another area for research?)

Sound symbolism

It has been axiomatic that there is no relationship between an object and the sound chosen to signify it. The label 'tree' could easily have been the label 'treb' with no loss of understanding.

Syllabic structure-

Syllables can be simple or more complex; the latter analysed as having a centre, plus a beginning (onset) and an ending (coda).

1) Minimum syllables are single vowels in isolation - words like 'are', 'or, err'

• Note: the surnames ORR , EMM.
  Supposition: Minimum syllable surnames seem to be comparatively rare

2) Syllables with an onset (i.e.more than just silence precedes the centre)- 'bar', 'key', more

• Examples: the surnames KEE, MAWER

3)

If you came to this page directly, then please access
Modern British Surname Studies
Last revised: August 10, 2008.