Hoare Genealogy 01: Origins of the Hore and Hoare Families

A Progress Report

By David G Hoare

Created: 5 June 1999; last updated 8 February 2017

This is an informal progress report on my work investigating the origins and development of the Hore, Hoare, Hoar etc family or families, written in June 1999 for family and friends who may share the same interests.

Contents:

1. Introduction and summary.

2. The contemporary distribution of Hoare households.

3. The historical and geographical origins of Hoare families.

4. Extracting provisional genealogies from geographical clusters of events.

5. Conclusions and prospects.

1. Introduction and summary.

My approach has been shaped by several considerations. The first was the existence of Captain Edward Hoare’s ‘Hore and Hoare’ genealogy of 1883 (H&H) and Burke’s published Hoare genealogies (Annabella, Gately Hall, Luscombe and Stourton) which extend the majority of the lines of the H&H genealogy (including my own) up to the 1960's. For me, this pre-empted the traditional approach of constructing a personal Hoare genealogy. The second was a fascinating talk by Prof. David Hey of Sheffield University in which he used the mapping of the geographical distribution of records of birth, marriage, death, etc for a particular surname to explore the origin or origins of that surname, and its development over the passage of time. The third was a personal pleasure in using computers to solve problems, coupled to the increasing availability of genealogical data in computer readable form. Particularly significant for me was the Church of Latter Day Saints’ International Genealogical Index (LDS IGI), held on computer from which one can ‘download’ records of births, christenings and marriages (‘events’), transcribed largely from parish records and bishops’ transcripts principally in the period 1550-1850.

I decided therefore to reverse the traditional ‘family history’ approach of starting from an individual and working outwards to an ever widening pedigree, and instead to start from the mass of genealogical data (‘events’ such as birth, marriage, etc), trying to detect the broad origins and development of particular families and ultimately move closer to reveal these as genealogies composed of identifiable individuals. In particular, I was interested to discover whether the majority of living British and Irish people named Hore, Hoare or Hoar could be related through the H&H and Burke’s genealogies, which during my life I had rather unthinkingly assumed to be the case. Since there is no mention of Hoare in the Register of One Name Studies, this seemed a relatively virgin field.

I started therefore by exploring the contemporary distribution of Hoare households in Britain. A geographical mapping, backed up by simple statistical tests, confirmed how heavily the Hoare name is concentrated towards the south and the south midlands of England, with apparent concentrations in particular towns or cities such as Oxford and Portsmouth (see map 1). There appeared to be about ten times as many Hoare households as could be reasonably accounted for as descendants of the H&H and Burke’s genealogies.

A mapping of the Hoare/etc IGI data, grouped into 50 year intervals over the range 1500-1899, has revealed three major developing clusters (see map 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850; note - a few events mapped way out to sea represent wrongly entered map coordinates, subsequently corrected - events just off the coast usually correspond to coastal towns just within the relevant map square). One cluster appeared to originate in Devon, just east of Exeter (and therefore close to Chagford which H&H identifies as the origin of its genealogy) but also at Axminster and Plymouth and spreading into Cornwall. A second is centred around Portsmouth and Southampton. A third developed in an arc of the Home Counties north west of London, stretching between Reading and Bedford (and including Green’s Norton and Walton which have been identified as the origins of other more recent versions of the H&H genealogy). There was also a smaller cluster at Gloucester. In Ireland, the great majority of records lie in a band stretching westward from Cork into County Kerry, and then northward to Kilarney and Tralee (see map 12). They are largely confined to the period 1650-1899, and coincide generally with the land grants to the Hoare family following Cromwell’s invasion in the 1650's.

I also explored the geographical distribution of Orr/etc (in England only) and of Hawe/Haw as possible spelling variants of Hoare/etc, but found virtually no geographical overlap (see map 10, map 11). There is a cluster of Orr, which is principally a Scottish name, around Newcastle-upon-Tyne, and a significant cluster of Ore/Oare around Telford in Shropshire with an apparent offshoot between Leicester and Rugby. Hawe/Haw is strongly centred around Thirsk in North Yorkshire, with large clusters near Boston in Lincolnshire and at Solihull, and a smaller cluster near Barrow in Furness.

I have investigated the cluster of Hoare/etc centred on Bovey Tracey (just west of Exeter and including Chagford) to see how far I can extract genealogies from the ‘events’ listed in the IGI. I have been able to assemble quite credible provisional genealogies, but containing a minority of weaker or speculative links (see sheet 2a, 2b, 2c). These provisional genealogies appear to provide a very valuable framework for confirmatory research into the actual parish registers, bishops’ transcripts and other relevant records, which is desirable in all cases, and essential for the weaker links.

2. The contemporary distribution of Hoare households.

I based my analysis of the contemporary distribution of Hoare households in Britain upon ‘The World Book of Hoares’ which I ordered from Halbert’s Family Heritage in 1990. It comprises a listing of about 3800 addresses of Hoare households in Britain (also others abroad); it was accurate when I checked the few addresses that I knew, but a minority of entries are evidently dubious.

I set out first to check whether the H&H and Burke’s Hoare pedigrees (which I had assembled as a computer database using the Pedigree program) could account for this number of Hoare households. Where a Burke’s genealogy extending to1966 existed, I estimated the maximum number of currently living Hoare members as equal to the number of those born with the Hoare name between 1886 and 1966 (i.e. during a recent 80 year span). Where a branch of the H&H genealogy had not been extended to 1966 by Burke, I estimated the number of current bearers of the Hoare name by applying an average fecundity deduced from the available Burke’s genealogies in the period 1886-1966. This yielded a very rough total estimate of 344 bearers of the Hoare name currently alive and descended from the H&H genealogy. It seemed inconceivable that neglected branches of this genealogy could increase this to account for the 3800 households in ‘The World Book of Hoares’. I concluded therefore either that the H&H genealogy must form only a part of a much larger single genealogy, or alternatively that the Hoare name may have arisen from several independent origins.

I mapped the 3800 Hoare households via their post codes. I first constructed a database containing just the forename and post code from the information provided for each household. The first element of the postcode (of which there were 1274 different codes in the data set) could be related via the postal address to a named town or city (of which there were 631). I constructed a database relating the names of these towns/cities to their map coordinates, and could then use these databases to print a map of all 3800 households (map 1).

Map 1 shows an evident preponderance of Hoare households in the southern part of Britain, and apparent clusters within particular smaller areas, such as London, Portsmouth, Bristol and Birmingham. Particularly with the latter clusters, I was concerned to check whether they simply reflected general areas of high population in large urban areas, or areas where the density of Hoare households is higher relative to the general population. I checked this my taking the populations of the 50 largest towns/cities in Britain (from an atlas) and matching them to their population of Hoare households (from the post codes corresponding to the named towns/cities). When these towns/cities were grouped into four geographical regions, application of the χ² statistical test showed clearly that the density of Hoare households per 100,000 population is not equally distributed between these regions; it declines from 21 in the south of England, to 9 in the midlands, 5 in the north and 1 in Scotland. The same test demonstrated unequal distribution between towns/cities within these regions; Hoare households seemed to concentrate particularly relative to the average regional density at Portsmouth (with 35 Hoare households/100,000), Oxford (44 households/100,000), Solihull (32 households/10,000) and Liverpool (14 households/100,000).

It is interesting that these regional and local differences of density of Hoare households exist and persist within the population as a whole, despite the high mobility of people in the twentieth century. I was interested to search for historical origins for these local concentrations.

3. The historical and geographical origins of Hoare families.

I was able to investigate the origins of the clusters of Hoare households by mapping the large amount of Hoare/etc genealogical data available from the International Genealogical Index (IGI), dividing it into batches corresponding to 50 year intervals over the range 1500-1899.

The IGI provides records of ‘events’ (principally birth, christening or marriage), each with its date and geographical location. It groups surnames phonetically so that Hore, Hoare and Hoar are grouped together as one name with a single phonetic spelling in the index. Each ‘event’ includes one to three names of ‘individuals’, either as child, parent or spouse (e.g. a birth ‘event’ will always name the child, and may add the names of one or both parents). However, a person may appear in several different events, first as child, then as spouse, and finally as parent, each with its own ‘individual’ reference, so there might be three or four times more ‘individual’ references in the data than people in the final genealogy. My initial Hoare/etc data set comprised about 14,000 events, and 39,000 ‘individual’ references, and additional data on Orr/etc (in England only) and Hawe/Haw names increased this to about 18,000 events and 49,000 ‘individual’ references. A rough estimate suggests that the 39,000 Hoare/etc ‘individual’ references might correspond to about 11,000 people in the genealogy, of whom about 7000-8000 would be born with the Hoare/etc name and the remainder would be spouses born with other surnames.

I have invested a fair amount of effort into writing utility programs to manipulate this large database. Some programs extract information from the records of events, such as place names, for which map coordinates can then be obtained. Some insert information into the record of each event; in this way each ‘individual’ record is assigned a unique reference number, and map coordinates are inserted to match the place name associated with the event. Others sort the events into separate batches on the basis of personal or place name, or date.

The essential steps in my processing of the IGI data are the insertion of a unique reference number for each ‘individual’ reference, the insertion of map coordinates for each place name (this required a database of 2700 place names with their map coordinates), and the sorting of the events by date into batches corresponding to 50 year periods (1500-1549, 1550-1599, etc). The Hawe/Haw events were also selected by name to separate them from other names such as Hawes/Haws/etc, which I considered to be too phonetically dissimilar to Hoare/etc to include. The Orr/etc events were selected by place to select only those occurring in England; this was to reduce the size of the data set. Orr is primarily a Scottish name with which I was only concerned only as a possible spelling variant of Hoare/etc which occurs almost entirely in England.

The maps of the Hoare/etc data (see map 1500, 1550, 1600, 1650, 1700, 1750, 1800, 1850) demonstrate clearly three principal clusters (described above in the Introduction and Summary), but they must be treated with some caution because the geographical coverage by the IGI is not complete and absence of data may obscure part or all of some clusters. The geographical locations of the historical Hoare/etc clusters certainly coincides convincingly with the present concentration of Hoare households in the south and south midlands of England. It is interesting also that Portsmouth and Oxford, which currently show an exceptionally high density if Hoare households coincide with similar historical clusters of IGI events.

The distributions of Orr/etc (in England only) and Hawe/Haw (map 10, map 11), described in the Introduction and Summary, are interestingly different from that of Hoare/etc. I had expected the similarity of the names to have resulted in considerable overlap as clerks had in the past often to fall back on phonetic spelling of names. This was evidently not the case, perhaps because there was some regional consensus on phonetic spelling.

The distribution of Hoare/etc in Ireland is in map 12, and was discussed above in the Introduction and Summary.

4. Extracting provisional genealogies from geographical clusters of events.

The essential element in constructing a genealogy is selecting and matching correctly from this large data set the events and their ‘individual’ references which identify a particular person’s birth or christening, marriage, and parenthood. An important assumption underlying my approach has been that, before modern times, most people were born, married and raised their families within the small compass of a few neighbouring parishes. I therefore wrote a utility program which allows the user to specify a range of dates, and a geographical centre and radius, and which in response provides a set of chronological listings of events within the neighbouring parishes or places falling within the user’s specified time period and geographical radius. An example is attached (Sheet 1) centred on Bovey Tracey with a radius of 25 km and between 1600 and 1625. Many families can be picked out at once with fair certainty. For example, George Hore married Suzan Connant at Hennock in 1609 and they had children Marie (1611), John (1616), Thomas (1617), John (1619) and Jonas (1622). The identity of George Hore as husband and father thus depends upon the matching of ‘individual’ records with reference numbers 11196, 22005, 16657, 30461, 16672 and 19713. Some matchings are of course more certain than others. Jonas’ birth cites both George and Susan as parents and so can be more confidently matched to George and Susan’s marriage than Marie’s, which cites only George (? just possibly a different George). George appears to have been born locally, but it would be harder to match George’s birth record to his marriage and parenthood if he had migrated some distance from his birthplace. However, the program can display likely candidates over a wide radius. The user can specify an increased radius of say 250 km, and then narrow the search by requesting the program to display only births and christenings, and only of people whose forenames start with ‘George’, and whose birth/christening date lies in the range 1569-1594 (assuming George was between 15 and 40 when he married).

In this pilot study 25 km around Bovey Tracey, it was possible to group practically all births and christenings which specified one or more parents into about 20 provisional genealogies. Several of these were just a single small family, but others were quite extensive and the largest, descended from Osmondus Hore, is attached as an example (see sheet 2a, 2b, 2c); George Hore referred to above is Osmondus’ second son. The genealogy lists, for each person, all the individual reference numbers that have been matched in establishing his/her identity, with any name variants. I try also to assign a rough level of confidence to these matchings, which is included on a scale 0-5. For example, conf2(child-spouse) indicates a fair level of confidence in matching the correct birth and marriage events - see entry for George Hore. A confidence level of 0 indicates that there are several possible options - see for example George Hore’s sixth child Susan who is shown married to Thomas Comyn. This is marked conf0(child-spouse), because there were three other Susans who might reasonably have been identified as Thomas’ wife (who are also listed). The full parish register entry for this marriage might contain more information which would resolve this type of uncertainty.

5. Conclusions and prospects.

I think this project is at a stage still more concerned with prospects than with conclusions. Yet it seems clear that present Hoare households are far too numerous to be offshoots of the Hore and Hoare genealogy alone. The mapping of the IGI data demonstrates clearly at least three major historical clusters of Hoare/etc families which appear to provide a broader base for the evolution of the present Hoare/etc population, and appear to correlate with their present geographical distribution. The existence of these clusters raises the question of their nature and their origins, and whether each corresponds to one or more independent origins of the Hoare/etc surname. The fact that the clusters appear to spring from a mere handful of people in the 1500-1549 period probably only reflects the fact that statutory parish records started only in 1538, rather than that the clusters originated in this period. My pilot study around Bovey Tracey suggests that the IGI data can be analysed to yield quite a detailed picture of these clusters, revealing individual families and their location in village and town communities, and possibly also some idea of the extent of personal mobility within and between these clusters. These provisional genealogies which can be derived from IGI data appear to provide a detailed and specific catalogue of questions to check against the possibly fuller information from the relevant parish registers and other records and sources, and also to identify specific localities where study of the local history may be particularly fruitful. Ultimately, the questions of the origins of names and genealogies can only be resolved at the level of the lives and movements of identifiable individuals, as has been shown by the debate about the origin of the H&H genealogy, but these solutions seem more likely to emerge when the framework of Hoare/etc clusters is better understood. But I am conscious that I have hardly started on this work, and that quite a different picture may emerge as I progress.

6. Methods.

I have used the Pedigree genealogical program (version 3.6J) as my principal tool for entering, exploring and manipulating genealogies. I downloaded the IGI data in PAF2.1‑2.31 format (not PAF3 format) and with only minor amendment this could be incorporated directly into Pedigree programs. My utility programs have been written in Borland Pascal v.7.0

Maps are prepared by incorporating the figures produced by my map_pop3 utility into a map grid produced and printed from the Corel QuattroPro (v.8) spreadsheet. This is scanned into a Corel PhotoPaint (v.8) graphics package, where it is merged with a previously traced and scanned coastal outline, and the composite image is printed.

The computing environment is a Windows 98 operating system, with a 350 MHz Pentium II processor, 128 Mbyte RAM and a 10.3 Gbyte hard disk.