00 seventeduhlentury he rate of decline was slow in comparison with that shown by the genetal population daring recent decades. 7. Save for epidemics and infectious discates, litle is known about causes of death prior t0 the nineteenth century. In catlier centuries, tuberculosis threatened children and young adulks far more than adults over 50 years of age. Smallpox, frequent among children, was a rae disease for young adults and. more or less absent among persons over 40, The diseates ro which older adults succumbed were mosly 20 illdefined that the reports are of ite use in tracing the changes overtime. better inferred by projecting backwards from the data for the present. Thus taking the quarter-century between 1931-3 and 2957-8, in which morcality ‘was very substantially reduced in England and Wales, the numbers of deaths registered for the great killers of our time, cardiovascular-renal diseases and ‘cancer (group (d) in Table 12), increased while the importance of all other causes (including tuberculosis, diabetes, anaemia, etc.) diminished considerably. ‘Taste 12 England and Wales, annual mortality per 1,000 persons by sex and age eats aS sis 195759 Are 554 @ cardionsubs wndeancer 509 67 59 <0 @) al ocher cscs oe 2 7 Fas Ae 5-64 (9 eardiovssuie and cancer an. RpbLon . Hilon . DV. Gta. Ce. Ei tule Cond breil BBB 6 EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE J. HAJNAL The uniqueness ofthe European patern |HE marriage patter of most of Harope as it existed for atleast two cen | tries up to 1940 was, so far as we can tel, unique or almost unique in the world, There is no known example of a population of non-European iviliztion which has had a similar pattern, ‘The distinctive marks ofthe ‘European pattern’ are (t) a high age at marriage and (2) a high proportion of people who never marry at all. The ‘European’ pattern pervaded the whole of Europe excepe for the eastern and south-eastern portion. [Lat us consider data for 1900. The European pattern extended over all of Europe? to the west of a line rumning roughly from Leningrad (as it is now called) to Trieste. The European pattern may be sen in Table 2 below. On the other hand, the countries east of our imaginary line are shown in Table 3. ‘Several of the Slav countries in Table 3 displayed quite a different marriage Tasiz x Selected Baropean countries in 1900: percentages single t selected ages (Single population as per cent of total popalation in age group) Slother exes “Mee Ween Cony joy is ase ee ae Aer Bec pte ES rere Belgem s 9 no on (9 eudiorscalr and cme eas ee es eee @ Slower exes es de "Bena Bape pro: ‘alga 8 on 3 4 3 : 8585 tas 8 Seca oe wo ‘9. In the more distant pas, itis ikely that the group (2) diseases played an even smaller role than in England in 1931-3. In addition, since in earlier cen turies the population contained a substantially smaller proportion of older ‘persons, the group (¢) diseases may well have made relatively litte impression on the minds of contemporary physicians. But it is these diseases, and the development of new prophylactic and therapeutic methods for dealing with them, which will determine the possibility of furcher increases in life expectancy in developed societies in the furure. [Notes Figures elate to veto of roo, Foe wousce we Table 3 Gi omens dt re sn Bysshe a bl frm subsaaty Eatope, he plea oan se more convent for the present pecs han they Beco aes he 190g 18 Was, powar data sc infloenced by Ce asa of he wat ‘which cated an abnor sper of wane ee $e wa mos convenient wt to have» stm fr the aea where th Baropan pate obtained sd Tiare le (wae ther so pomaley of minndctnaiog to we Baroy to Scot is ‘es. We award fo exelate Exten Europe from Europe and eaighe be thought nore secure {0 we crm ke "Western Europe and "Weren Eusopecs pacer However ace ts concept ‘bad co be refered oop feauenly, brevity ws ares sdraenge Borope Gur tied vse in Gir the are dak wit in many a Ratry of Europe. >102 BAINAL pattern fzom the European one. Let us call theirs the Eastern European pattern. ‘The volume of data in Tables 2 and 3 is hard to digest and, as a beginning, iis beter to selec a few countries for comparison. We may (Table t) contrast say Belgium with Bulgaria, or Sweden with Serbia. (Aliteration is 8 good a principle of sclection as any other} “Table 1, like the succeeding tables, shows the proportion of the population ‘who are sil single in cereain age groups. The numbers rentaining single at 45-49 may be taken to indicate the numbers who never marcy at all. In the European pattem (Belgium, Sweden) a substantial proportion remain single throughout lif, in the Fasten European countries almost none. The propor- tions single at 20-24 and 25-20 are indications of the age at marriage. Here also the contrasts clear. For example, by age 20-24 some three-quarters of women are sill single in che European pattern, while in Eastern Europe three-quarters are married inthis age group. Taste 2 Burope (except Ease Europe) around 1900: percentages single at selected ages (Single population as per cent of total population in age group) ew Won Country a ee poy 2539 ae To os = oe 8 3 agian % nm 8 4 Desmae B50 2 % @ 43 Filind Bow a 2 os Fence o oo Ss 8 oo Gexwany x 8 9 n © (Great Bria ho 2 pn 8 8 Holl & 8 8 pH Teeknd a & © mB eho’ nd % © ily fon @ % @ Norory on 7 ee Porogal rn o «2» Spain HO HO 80 sO 6) 20 ‘Sweden 2s on » 2» Swine x 8 6 ee ‘Nota: Poi bouodaies ws oF 190. ‘Dats bed on cee aken in 1900 and apor excep for Irland (891) and Hay (910). (By ‘ese vans in date ts pone o achieve condneacy inthe age groups wed exept for Spain) ‘Age groups for Spin) 235) 2630 (9 460 ‘SSnc ast Inecatona do Seige, Anse Inno de Stati, Vol. 1, at de Popoltion Europe, a Haye 15:6 ‘The reader may satisfy himself by the study of Tables 2 and 3 that this con~ trast is not due to an arbitrary selection of countries. Table 2 shows alP che ‘countries sharing the European pattem arranged in alphabetical order. Any of them would have done equally well for purposes of illustration, There are sany differences between the macriage patterns of various western European countries, but there is a distinc cleavage between any of them and the Eastern European pattern of Bosnia, Bulgaria, Romania, Rusia or Serbia (Table 3). * Some smal ras (saxembusg, Farce lands cz) have ben omited. EURODEAN MARRIAGE PATTERNS IN| PERSPECTIVE 193 ‘The cleavage is especially marked for women. For example fewer than $ per ceat of women remained single around their soth birthday in Eastern Europe, whereas in Table 2 the figare is nowhere below 10 per cent and often above 15 per cent. In the European pattem unmarried life for an adult woman was accepted as a normal {if perbaps exceptional) alternative to marviage. In Eastern ‘Europe this alternative sarcely existed. Taste 3 Easter Euroge around 1900: percentages single at selected ages (Single population as per cent of total population in age group) Dae ey “ew ome ‘Com wy eed 22D AH sexe Base Gree ee 3 eB + Honey 1900 2 5 eos ‘ Romint hop 6a 3 > 3 3 Bona to) BED 2O 6 20 Bugz goo Sh 3 Ps 2 1 USSRY ie st 3 > ‘ Sctbia yo 3 6 2 : Nor Age rouge: ( ai-ae, @ azae, @ 4150. Saou Sane a able 9, exept USSR. gues ken from United Nations, DemarapicYertok, soos, Tbe 6. There are, of course, intermediate possibilities berween the European pattern of Table 2 and che Eastem European situation—for example, in Hungary ‘and Greece. (The countries in Table 3 have been arranged in descending ordet of the figures in the first column.) The populations included within the bound~ aties of a sovereign state are not necessarily homogeneous. It would be a ‘worthwhile study to trace the variations on the fringes of the area where the European pattern prevails. Significant departures from the European pattern ‘ay probably be fond not only as one proceeds eastward but on the southem edge of Europe as well. Parts of southem Italy or Spain are more like Greece than like Belgiam or Sweden. How far would 2 more recent date than 1900 have yielded a different picture? Any date up to about 1940 would have left unchanged the basic con- trast between the general European pattem and Eastern Europe. Indeed most of the figures for individual countries would have been substantially the same, ‘except for Eastem Europe where the revision of frontiers fier the 914-18 War altered drastically the composition of the territories for which statistics are available. Also in some areas ‘modemization’ may have brought about some ‘hi away from the traditional and towards the European pattern. 1 For taepesn Rain 157, the eet ng wre ow Wener wo 8 3 oo ‘ 3 + lotermeditepastena have ao bon characte of the Unlied Sees, ‘TToe major eacepon i France where thee wat comidcable redaction inthe percmsages single among women ovr the Ss tneedacdes of this ceatury. Indeed the reduction Wa lead Fern Sele tof eer cry. Range ntibrepe die rinse fe iy ae eens to have aatclpted ats leur pace changes Which euewhece in Europe Wace rake plc much fats Soe Bourges ich, bow, Pep. #85-104 AJNAL In the lase ewo decades much of Burope has experienced something like a revolution in marriage habits. People mafry more and earlier than in former days. The percentages remaining single, cspecially among women under 30, ae now far lower than in the earliee ‘European patter’ illustrated in Table 2 and the numbers remaining unmarried throughout Ife are being greatly reduced and may in several countries fall below 5 per cent in the near future, ‘The ‘European pattem’ seems to be disappearing, ‘Non-European civilizations ar like Eastern Europe, or more so. Percentages single are very low by European standards, at lent for women; in the age ‘group 20-24 there are often fewer than 20 per cent single. (For men the contrast with Europe is less clear.) Very fow women remain single throughout life (not infrequently z per cent of les) and for a man to remain a bachelor is not much ‘more common. Figures for many of the larger countries in Africa and Asia are reproduced in Table 4. The countries ae listed roughly in order from west to cast. Tass 4 Aftca end Asia: percentages single at selected ages (Single population as per cent of total population in age group) Daeeh Contry mt 20g Nooee Nici pe 2 oes ges (Modem) it on a om 2 “Timia Gndigenoas ‘populsion) 546 a zoo + aye 147 CT 2» 2 ‘Mozmbique 1950 st yor os Manzias 35 2 eB ops ‘Tukey 3935 ° 6 3 Jeon ne Pakieas) 3932 3s stare. Cera. as fe Bon 3 ‘Thaland ser 6 eon 3 ‘Malays (Malay) 3307 + 7 2 8 Formos. ane = B48 Korea wie 3 2 te Jpn 1920 m m9 3 Note: All gers elt to terres 5 oF the dates sated Sources Morocco, Alga, Tuna, Maueias Usted. Nation, Donogephie Yorleok 1935, Table Mozambigue—Portpl, Provincia de Mocambiqne, Repariso Tecnica de eatsicn Reamer geal de populjte em 1950. Vol, Popalacio nao cvliado" (Lowengo Marque, “Table 1959) pp 1218 ‘Other coustio-J. Hija, ‘The Macage Boom, Population Ides, Vol 39, No.2 (1983). ‘There are considerable dificultes about comparing figures such as those in ‘Table 4 with European data. In the first place the family as an institution has different characteristics in different cultures and it is not possible to define ‘marviage’ in a unique way for statistical purposes in all counties. Clearly the term ‘married’ has a different meaning in a country such as India where child smatriage is traditional from that which it has in a European setting. For the BUROPRAN MARRIAGE PATTERNS IN PERSPECTIVE 105 purposes of this paper marriage means roughly entry into a union which is regarded as appropriate for the bearing and rearing of childcen in the society in question. Such a definition will be meaningful in very many societies, For many areas the available data on marital status unforeunately may not represent anything even remotely corresponding to the definition just sketched. For India it would clearly be preferable ro have data where only those were classified as married who had completed the second marriage ceremony (‘gata ceremony) after which consummation takes place. Even after that it seems customary in some areas for the gel to rerum to her parents’ home for months or even more than a-year so that she only joins her husband permanently some time later. In Indian data people are counted 2s married who, for our purposes, should not be so clasfied. The opposite dificulty occurs in many cases. Some- tinses offical statistics count as married only persons who have been declared married by the civil or ecclesiastical authoritics, whereas the people themselves take little notice of the requirements ofthese authorities. In some societies there are several accepted forms of marital union, only some of which are clasified as marriages in the census. The precise scope of census classifications even 28 intended by those in charge of planning the census is offen unclear; unimagina tive imitations of Buropean classifications have often been used in societies where these were inapplicable, How the definitions were applied asthe entumerae tion was carried out is even more dubious. ‘In many instances (e.g. Ceylon and Japan) the proportions of persons single if the tre facts were known would undoubedly be even iower than those given in Table 4. Marital status data for Latin American countries and the Caribbean area have been langely useles because so many people who, on out type of definition, ought to be counted as married have been treated in the statistics as unmarricd.’ More accurate and meaningfal data chan are yielded by the official censuses arc sometimes available in special studies (c.g, the Trinidad survey described by Braithwaite and Roberts). Information on the mumbers of women who have bome children may also be used in several countries to show that the marriage pattern cannot be of the European type.* ‘A further dificulty is that figures on age distribution are known to be highly inaccurate in many countries. What is worse, information on age and ‘marital status may not be independent, ie. chece may be a tendency for census 1 Thc is recent review ofthe svalble formation by Moras. (Det of works metoned may be found inthe li of rekrenee a he end of is eee) “" cllcson of dts of ths kind ay be found in United Nations, Doneepic Yearbook 1935, Table. The securay and eaninginnet of aval satus aa for Baopera Cooma of es, otaiwsysabovenuptaon foreatpl if ome lon wast kes to scouar che feos Beers belowerhas he pvenn Tabs rts Pornnceentxoney Sra rate ‘gnacy rte snd relent the mrarre co comer cians (ocelar axeaps) sagt tc the cet might be abwantal Feely data may be ete to sow ennecdng SiGe rei lanl wie pe beech Boop Sg aon, For Bropen erie the fey data shuld rele os pid bear he widtpread we of ray ‘Te fillowing comparison of age peti ate ed on lb, legate sod eaty th ferences between Secen aud Bulga i he proporuons of womtes fing ont fats eage ees (hope opt desea ah pecanage cf wlows rete Eo aceon stixerpetig ach ces) ery rats pe 1000 woe fl mar odo) Asegoe Te aoe asa Bal os 4 * mo (Daa om Kita, The mere of popaon govt, 135)106 ASNAL cenumetators to put down a woman as being above a certain age if he is married, but to treat her 2s younger ifshe is unmarried. The fact that most of the figures in Table 4 are for recent dates must also be remembered. For our present purpose the proper contrast would be provided by figures which described non-European countries as they were before their social structure began to be modified by changes derived from Europe. If we had statistics going further back in time (e.g for nineteenth-century Japan) the contrast with Europe might well be greater in several caves chan Table 4 suggests. ‘No survey of the information for small pre-literate societies appears to ‘exist, However, it would seem that in the large majority of these, marriage, for ‘women at any tate, is early and unis ‘When all the qualifications about the data have been made, there can be no doubt that out original generalization remains. The European marriage pattern is unique for all ltge popalations for which data exise or reasonable surmises can be made (eg. it may be surmised that those Chinese populations for which vwe have no data are similar to those for which figures exist). Europeans have sarried very much later than others and far more of them have remained un married throughout life. In non-European civilizations there are scarcely any single women over 25. It is not, of course, intended to suggest chat non ‘European civilizations do not show wide variations in the pattern of their sarriage rates. But all the varieties chat exis are separated by a distinct gap from the European pattern, The eighteenth century ‘The zason why Foar Womea out of sx do not bea citéen every yest is that they cannot marry becawe of the dtcouragements and Aificulics in ete way. Riausso Camiov: Esa sur nature di commerce en géésal (1755) Ifthe European marriage patter is unique, it s natural to ask, ‘when did it aise?” Curiously enough, this question seems scarcely to have been asked, let alone adequately answered. The suggestion has occasionally been made that late marriage is characteristic of urban-industral societies while agricultural countries have early marriage. The suggestion is certainly unfounded. Figh- teenth-century Scandinavia can hardly be described as urban or industrialized. The question about the origins of the specifically European marriage pattem ought to be answered by historians well versed in Burope's economic and tocal history back into the Middle Ages, as well as experienced in handling such statistical material as can be reconstructed for petiods earlier than the sightecnth century, A demographer accustomed only to deal with modem data cannot go very fr. ‘What follows is really only a survey of the most easily accessible territory. ‘There may well be much more material, even of a seatstcal nature, in places where a demographer would not be able to look for it. ‘The frse sep is to see how far back into the past the specifically European pattern can be traced. The type of data used in Tables 1-4 unfortunately does UROPEAN MARBIACE PATTERNS IN PERSPECTIVE 107 not go very far. Censuses with cross-tabulations of age and marital status do not go back even in Europe before about the middle of the nineteenth century ‘except for some Scandinavian materials, For Norway and Denmark we ox go back to the end of the eighteenth century, and for Sweden the cross-clasifica- tion of age and marital status has been reconstructed back to 1750. (Eor the Scandinavian data, and nineteenth-century data for other parts of Europe, see Hana (Pop. Index, 1953).) There is alo some information from mariage registration. Marriage in Scandinavia in the eighteenth century may well have occurred earlier and there may have been fewer remaining single throughout life than in the ninetceath century, but there can be no question that marriage habits conformed to the European pettern and not to the pateem common 0 the rest of the world. Nor indeed would one expect anything else, Historians on the whole appear to have taken it for granted that marriage pattems in Scandinavia and indeed throughout Europe in the eighteenth century were of the sume general type as later. The changes in the marriage habits which have been looked for hhave been changes in the opposite direction, namely towards more and earlier ‘marriage a8 a possible explanation for the spurt in population. growth, ‘Much more information could probably be extracted from the early censuses or enumerations, When the European pattern prevails this fact will usually be clear from any data on the distribucion of the population by marital status, even without cross-clasfication by age, since the gap between the marital status distribution characteristic of Europe and thote of other civiliza~ sonsis substantcl, Of course, ifno cross tabulation of marital by ageis availabe, i will not be possible to deduce whether the proportion remaining single throughout life is high or whether it is only late marriage which is responsible for a high percentage of unmarried persons. Unforcunately most of the published information before 1800 on distibu- tions of populations by marital status relates to cities. Cities, however, are far from being closed populations. The dstibuciem of theit populations by marital status is influenced not only by their marriage habits but also by migration. European citics in particular have frequently shown signs of this by having a surplus of unmarried persons and especially women (ia former days many of these were servants). This ‘occurs throughout the centuries back to the Middle Ages. The tight inference to draw from a high proportion of single women in a cry i often not so muck that urban life discourages marriages but that cities provide opportunities for single women to earn a living and single women, therefore, go to live there. However, some information is availble from a number of counts covering ‘raral populations in the eighteenth centary. A review of this material has been placed in an Appendix to this paper. All of it, as far as it goes, suggests the general conclusion that the European patter originated before che eighteenth century. For studying martiage in the eighteenth century, we are not confined to data on the distbution of the population by marital status, The distribution of marriages by age, and the mean age at marriage calculated from it, provide ppethaps the most natural approach. We may begin with the series for Veniceei by Belteami.* These figures ace mean ages at marriage forall marriages ‘without distinction between first and later marriages. rors 209 aa ines sre a8 toe oe aos iNet es Be Hoy a ato To interpret figures on mean age at martiage'® we must inquire what mean ages should be regarded as characteristic of the European pattern, and what levels would indicate a non-European pattern. There is little direct information on the distribution of marriages by age for countries where the marriage pattern is non-European. Such countries in many cases have no system of sarriage registration at all or else the registration system covers only an in- adequate portion of the marital unions, as explained earlier, Even where the registration system covers the bulk of marriages, statistics of age are often tmn- reliable since many people do not know their ages. However, information on the age at first marriage" may be obtained by indirect calculation from the proportions of single persons in successive age groups. Calculations of this sort suggest that a non-European pattern implies chat the mean age for marriages of single women is below 21. According to the European pattern the mean age for the marriages of single women must be above 23, and has in general been above 24. leis much more dificult to specify limits forthe overall meanage of marriage of women, ie. ifthe marriages of widows and divorced women are included. (ln European populations in the cightcenth century and earlier, there are forall practical purposes no marriages of divorced persons) If we assuine that the mean age at frst marriage is at most 2-0 years and make rather extreme assumptions about the proportion of marriages which are remarriages and. about the mean age of the bride at such marriages! it * See Bebramk p. 181. ‘The orga dination: (by single years of zg) ace given in the Ape “abet marge ir 28) a Beta book rte ge op a9 sot 40The crude ie, the average of the which scr rel Sm es ie el es See sat yas aap eee IES See eae ene Uae ee ee Bice ey one neenions wee mage tee ae ieee Sine beeen clare antee seid nee weaee erceeeat ee aca ae nga nm ng oe eee lessee Ernie fons tg meee age momen eas Ee eae oe ee ‘“BThe ewo Gena (he grap of weecage sad the ets ope a Femang) an nat eget Sewer meee ied eat Siete higher nce ae af ces Ws wie E ple Breioe manage thse higher perctange of marge in bow European wees an ia Earope: br thes aime fcr lower the aveage ape af which widows tomarry by compara increases he propor of remains, bur veal oomnen EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE 109) seems likely thet the overall mean age at marriage in non-European socicties could hardly exceed 25 years. The subject needs investigation. Probably the overall average age at marriage of women would in fact almost always be below 24 and indeed usually below 23 years in societies ex hibiting 2 ‘non-Eoropean’ marriage pattern, As an actual example, the following data for Serbia may be quoted. Mean age at maricge Men age at mariage ‘fap ones ‘all women) 1886-95 200 217 psI095 27 24 In societies with a European pattern the overall mean age at marriage of women is sometimes at low as 24°5 (Italy 1913-14), but ustally a good deal higher. cet ee We cee vez eae ae they are clearly of European type. The number of marriages in which the age of the spouses was not given is remarkably low for eighteenth-century Venice. However, in parish registers of that period, statements of age are generally not given or are too incomplete to be of much wse. There are, therefore, very few studies making use of age statements at marriage registration. According to one study! the age of the participants was stated in the records of 63 of 83 marriages celebrated in 1664-9 and 1693-5 in the village of Someren in the Netherlands. The average age of the bridegroom was 27 years 4 months, chat of the bride 26 years 8 months. ‘Another Dutch study (by van Nierop) takes us back another century. It is based on the registers (vil, not ecclesiastical) of the city of Amsterdam and ‘covers 11,597 marriages contracted in 1578-1601. In 9,247 marviages the bride~ groom was a bachelor and 8,052 of them stated their ages, though only 4,664 did so without adding some indication that the information was approximate. ‘There were no bridegrooms under 18, The percentage distribution of the stated ages (including approximate ones) was not unlike the distribution of the ‘ages of bachelor bridegrooms in England and Wales in 1891-95: Ansirton England and Weles ‘s7esbes es rey 2 A 2024 a @ 3e9 aM 3 seandover B " remy a+ younger vege age dan nies Within mend Earopean cree He ee eg ae ee hs peel ser cede a pe cad 5 tuly been below 15 er cent 40 Seams 3 ridows marys (98 average age the tan age of widowe marrying in Poplnd and Wes a the beginning of tie century) we ‘ke ao po cont for the proportion of remarige and fo forthe serage age 3 Temarige We ‘besa a overall mean age at marriage of 2, ce aixoeeyoxosmsys prosiatly the uve figure rested if we ice the proportion of remarriage to 2$ ek et teenage se ot hace Sch mag ony joe | {he proportion of rameiages 9 per cea, but ce average age a eauage 94 Yur ee BT havent seen the original paper by Susen. The deals quted are Sema Mab (Val. 1, 237) Male monumental werk provided invaluable gue to the sourse Zoe the peparton (Se present paper.0 ANAL Unforeanately van Nierop made no analysis of information about the brides; she felt the bridegrooms were more inceresting. In general, however, in marriage registers of the eighteenth century and carlic, statements of age are not given oF are too inaccurate to be of much use. ‘This dificulty can be overcome by the laborious procedure of matching cach marriage certificate with the certificates of the spouses so that their ‘ages can be determined directly from thei date of birth, The earliest seudy in ‘which statistics on age at marriage were obtained by this method may well have been that published by Roller in 1907. It rclatcs to the stall town of Durlach in Bavaria (Germany). The mean ages at first marriage were as follows: Mow Women yo120 287 368 itso are a4 Histo 28 356 78-0 56 ast For comparison it may be added that the average age a frst marriage in the whole of Bavaria in 1896-7 was 27+4 for men and 24-9 for women, There is ‘hus no substantial change from the eighteenth century. In the pioneer study by Morrell (1935) of the registers of ewo English parishes (North Elmham in Norfolk and Wedmore in Somerset) ages at aarriage were obtained by the matching of martiage and baptismal records, ‘This stay i also remarkable in going back to the sixteenth century, The overall ‘mean ages at marriage (including remarriages) were as follows: Norte Bian Wednave 561-1608, reas Men 276 279 Women 2s 208 ‘The numbers of marriages upon which the averages are based is not give tot irscems lily from other information that thie number exceeded "0 in each parish, It is not clear whether it was posible to establish the age of the spouses forall the marriages taking place in the stated periods. In recent years the technique of matching records bas beea applied in a number of studies. Such work can be greatly asisted by modern data processing equipment and is likely to be undertaken on an increasing scale, so that, in a few years, much more information on the age at in the seventeenth and eighteenth centuries should be available. The series given in Table 5 relates ‘to an Italian village and covers a period of 2} centuries Is the product of an extensive study whose focus is on human genetics. The project is under the dizection of L. L. Cavalli-Sforza who has very kindly made available the data seproduced in Table 5, It will be seen that the mean ages at marriage fuctaate (pardy on account of small numbers), but no trend appears. ‘The matching of parish records was also the technique employed in the well-known study of the Prench village of Crulai in Normandy published by Gautier and Henry in 1958. No change appears within the period covered, " Thave sot my xn Rolls uly. The Sgces on ge a manage ate aken fom Mom ‘ae, Bethel, p 4. Mala (VT, p. 36% no) pga ta hey ee rms ph Ml (VL , -28h oe gan tt they were cee 9 ce BUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE a Tats s Marriages in village of Riana (Parma diocese), Italy ‘Aveage cae of Avge age of Yea ‘ido tie SE. 1a » ascend io » waze wae 33099 w Borst Gorter Heo9 2 panes yoataa 385099 8 sages arsire [Not Resced to mariages whez bec bdo and bridegroom are bor in the lage. ‘The table refs eo all mrcape Gacading remarriage). SIE mesurandard ror. namely 1674-1742. The mean age at first marriage was 26+6 years for men and 2asrr years for women. For another French parish at a somewhat later period (5760-90) Girard obtained mean ages at first marriage of 27°4 for men and 26-2 for women, These figures compare with 28:3 and 2471 respectively forthe whole of France in 1851-5. We may also refer to'a figure of 2475 years for the mean age at fst marriage for women in Pars in the eighteenth centary. Bourgeoit- Pichat (1051) states that it is based on ‘recherches directes dans regisres de Vérat 3 Paris’ Some very interesting information on the average age at marriage in seventeenth- and eighteenth. England is given in J. D. Chambers’ ‘monograph on The Vale of Trent (1957)27 The basic data are not parish registers but ‘allegations’ of marriage for Gloucestershire and ‘certificates’ of marriage for Nottinghamshire. Only thote who wanted a quick or quiet wedding, ‘without the formalities of the usual procedure, are included end the poorer classes are under-represented. We are not told what percentage of all marriages is covered and the accuracy of age statement is noc discussed. Presumably te- ‘marriages are included, not only frst marriages. Chambers uses medians not means. Occupational groups are distinguished. This stady gives by far the carlest significant information on differences between the marriage habits of social groups.#® The Gloucestershire data go back to 1637. The median ages at marriage show no trends over time, They could come very well from nine- teenth-century or twentieth-cenrury European dsta. ‘Some further mean ages at fist marriage may be added from an unpublished © Theta aed on 272 manages facile ad au manag of oes epi Thee cage arte dae eich fake spt an iene epee, Beata ea hence bad ca wa ae faa Gea cca ESSE Sarasa PS pets po a aula ne cape EASE Ela i gee det noe aa Dahl Se Leyes ae oe Eoin dan ine ove hobptasmmna SE gs calc cases Rahs Sages a ile Pa hptenes ta sin nde ipl Rede ts aie Re Rae s espero SStaRp areas Soeur peat tne ear pra mitigeefones ir Paks Sew, uel ae PaoeDw cep 2 Ute ae Sa ane wal, selang to Bal a Yly 7, wa pb inane Seal Ge Be as a0.im ANAL paper by Paul Deprez® These are based on genealogies of Flemish rural families and subject to posible selection bias, perhaps in favour of the inclusion of those who left more olfipring, Such a tendency would presumably lead t0 underestimating the age at marziage, No information is available on the repre- sentativeness of the genealogies in regard to the social composition of the population. Here are the mean ages at frst marriage of persons born in the periods stated: Ntesy 70019 sassy asas——asheag Male 253 259 ae oer Female 3D 262 ast ast as Another type of data sometimes available for the eighteenth century is the classification of deaths by age and marital satus. One way of using this ine formation is to compute the percentage of deaths of women dying over say $0 years of age. This gives an estimate of the percentage permanently remaining ‘unmarried. Inthe study of the French village of Crulai, which has already been mentioned, there were in the period 1730-1800 224 burials of women over 50. 198 were of married or widowed women, 4 were single and in 22 cases the ‘rita status cannot be determined, The percentage of single women among ‘women dying over 50 may thus have been as low as 2 or as high as 116 (accord ing to whether we assume that all or none of the 22 were single). More likely the truth lies somewhere in between,® In any case the percentage seems to have ‘been low by ninereenth- or twentieth-century standards even for France (which had the lowest proportion in Europe of women remaining single throughout life) Apare from the possiblity of accidental variation, itis necesary to keep jn mind the effect of migration. Some women from Crulai may have become nuns and died in 2 convent away from the parish. Other single women may hhave emigrated to a city, for example a servants. The distribution by mirital starns and age of deaths in 1715-44 in the parish of Saint-Sulpice in Paris was published by Deparcieux (1746). Of the women dying over 50, 15 per cent ‘were single and of the men no fewer than 20 per cent; Deparcieux remarks that there were many servants in the district. For Pomerania the distribution of che deaths of persons over 25 by marital status is recorded by Stissmileh® for che nine years 1748-56, There were in all lover 40,000 deaths of persons over 25. 10 per cent of the female deaths and 15 per cent of the male deaths were of single persons Since the lower age imit haere is 25, so that some of the deaths took place at ages when marriage is not Yet too rare, these percentages are rather low by later European standards. However, they are far too high (especially inthe ease of women) for any non Baropean population. To sum up, the European pattern in age at marriage can be traced back in ‘many countris tothe first half of che eighteenth century or even earlier. There asin Bor terme Nori, Deca sate ey Dee Similar eas to those of Cut on the women dying single were obtained fom another ie the cn ofc eins san by Cad pent Apes tas BUROPEAN MARRIAGE PATTERNS IN PERSFECTIVE, m3 is no record anywhere of a non-European early age at marriage. (There is some suggestion in the data that fewer people remained single throughout life in the eighteenth century than afier 1850.) The aristocracy ‘The unmarried Ladies and Geadlemen inthis City, of modecateFor= tunes, which are the great Bull, are unable to support the Pispence eben wih ay Macon Sey. ert, see (Calbacy; Each Sex compensating ill, asi ons Gomer Minna: Ouaridanca be Geen a Present State of the City of Landon. ((751) Before the later half of the seventeenth century, there is almost no statistical evidence on marriage, at least so far as unselected data (ie. covering all in- Ibabitants of an arca) are concerned. A continuous record going back earlcr is available for slected upper class groups, and above all the aristocracy. Table 6 gives data from a study of the British peerage by Hollingsworth (19$7). The figures relate to the legitimate children of kings, queens, dukes and duchess, In genealogical records of this type it is possible to trace the vital events (marriage, birth of children, death) ‘occurring throughout an individual's life, Its most uteful to analyse the data by “cohor’, i.e. to consider individuals born in successive periods, and stady cheit marriages (s against the more method of stadying the manages occurring in successive periods). It is thus posible to consider the numbers remaining single of the survivors at successive ages. Two ages have been selected for Table 6, ‘The story told by Table 6 is a remarkably clear one. The frst two lines represent a marriage pattern quite different fom the later “European” one, but Taste 6 British peerage study oO x Wome Peat of Percent al sing ot Parcs gle t ee a o o 3° iene = > @ 7 ties 2 4 ss é iow 3 3 % 2 men a 2 Fd 4 ryott9 we 3 ® 3 rho ke x & 3 i en Source: Hollingsworth (199) p14. Themumbes on which penentage aad average age at mange were buted ac nt given. Ar bch hse somo ve ben between 129 and So of ach tex in oe cohort, ba of cous dhe numberof marriages was male, xpecsll nthe exly cor, 2 tens been suggested by Connell (150) chat early mariage wat characertic of eighnccah- cxotucy Isang. A speci sh caer age pattern Geveloped inte eendes afc ie oaiae Se ney te sod Right percentages runing amatred, Connclstheoey tat tere fad Pere ‘Eis Towards ce utinge it the decades teers the ee b wmponel Sea ‘ridene wou conideting. Even before the fomine he age at manage’ Leland ad howeeee rebate ing Nov We Byes cumcrs thom bye pene age ‘ecorded atthe ens 4 (See Hajal, ‘The Mariage boom’ (3953) 9-95, Nowe) For Gonaets views below, Pare Mp a23- f 13 Stebel, Part p34,mm magwar something like that found, for example, in Bulgaria, at least so far as age at marriage is concerned. The population remaining permanently single is higher than in non-Westem societies. From the third line of the table onwards the picture isa typically "European’ one. ‘The mean ages at frst marriage show the same drastic change* as we pass from the Middle Ages to the eighteenth century. Peco of bith rys0-re79 s48ou679 1680729 173079 THEI 8079 Mes ae aes as BS gos 0. ‘Women mt saya ‘We tur now to the records of what should, perhaps, be termed a republican aristocracy, the ruling families of Geneva. Their demographic. charactecisties have been studied ina monograph by Henry (1956). The figures for the earliest ‘groups (especially among women) seem to show craces of a former ‘non- European’ marriage patter. As Table 7 shows, the change from an earlier to ‘a Europem pattem may have occurred at roughly the same time as among the members of the British peerage; but too precise a comparison should not be attempted. ‘Taste 7 Ruling families of Geneva SRE nessa fet Eat ew Peete Sees eS Peete seus titi Peller in his study of the European aristocracies* analysed his material in a manner less suitable for our present purpose. His data suggest that in the continental European aristocracies a ‘non-European’ pattern persisted far longer than in the British peerage. The families studied by Peller retained a position of i Sec BREST args ne oe a wee SE hove ptr. a BUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE ms substantial feudal privileges and 2 distinc: manner of life far longer than their British counterparts. The very highest layers of the French aristocracy, the “Ducs et Pairs, also preserved a pattem of very early marciage, distinct from the customs of the majority of the French population, right into the eighteenth century, as recently shown by Henry and Levy. Indeed che comparison with British conditions was noted by at least one contemporary observer, the son and heir of the Duc de la Rochefoucauld who vised England in 1784 at che age of x8, His charmingly romantic view of masriage in England deserves quotation: “Husband and wife are always together and share the same society. It is ‘most uncommon to see one without the other. .. . They always look per fecily harmonious; the wife expecially looks so contented that it always gives se pleasure. ... 1am not sure whether having to live constantly with one's ‘wife makes it necesary to marry at @ much later age, but I am inclined to think ¢o. In England, to have 2 wife whom you don't care for mutt make life a misery. An Englishman, therefore, makes a greater effort to get t0 know his bride before marriage; she has the same desire and 1 believe that this is why marriage before the age of twenty-five or twenty-eight is rare. Perhaps another reason for this is because it is usual to set up house imme lately after marrage. The young couple never stay with their parents... “English husbands have an advantage over us of which they sometimes avail themselves, namely divorce’ ‘The analysis of the genealogies of some families in Wiittemberg (Germany) prescated by Riimelin (1926) can also be included in chis section. The occupa- ‘tons recorded show that the families involved belonged largely to the more prosperous urban sections of society (the agriculmral population is hardly represented). The selection efeess which may bias data collected ‘on account of the interest of descendants (for example, selection in favour of families which have numecous offpring) were mentioned above. ‘The precise nature of the records is not described in Riimelin’s paper and Taste 8 Age at fst marriage according to genealogical records of « family i, Waartenberg (Gemnany) = Won Ne. Mean age Novof Mean ae Cony oe at marisge ut aware sah = 3s * 1 rh Be 3 am aot eb 2 29 ° 260 3th = nt % 353 ‘Source: Rumelin (930), Tables 38 and 128 (Fanly ED. ‘Thee is no explaration why there ate so many mote cates for computing the male ag than far he female ages * Frnqa do Rothfoacld, Méanges av Page. The goctation is fom an exact teased in Franca M. Wibo, Sarg ind, London (Longmans, Gren 0d C2), 1955 fe6 BANAL there is no discussion of completeness or accuracy. Data from the one group of records which includes material on women’s age at marriage going back to the sixteenth century are summarized in Table 8. The men in this genealogy were largely salaried professionals, in particular civil servants and clergymen. ‘There are hints ia Table 8 of a non-European marviage pattern in the earlier crac No rcihs cn be ached hi at, bu fog records ofthe sed by Riimelin are sill in existence” they might be promising material fbr an amigas ox modern line areas It's, of course, entzely possible that, before the modem era, the ais and the upper classes generally married much earlier than the bulk of the ‘population. (In che nineteenth century, the aristocracy, atleast in England and Scandinavia, married late and less feequently than the population as 2 whole) ‘We now consider the scanty data available for unselected populations in the Middle Ages. The Middle Ages (Good Sis ince Iwas owelve, some years ago, (hanks be to God that I am sil alive) Of husbands ae the churchdoce Pee bad five. Cauaveen: Witt of Bath's prologue (Modem oro by Won We) A few urban administrative documents (mainly taxation lists) which peemic a marital status classification of the adult population are available even for the fourreenth and fifteenth centutics, The marital status classification maust be auade on the basis of indications in the original records which were intended for other purposes (eg. taxation) and the indications of marital starus are therefore more or less incomplete (cg. in the case of servants or adult childeen living in their parents’ home). Moreover, the age distinction between children and adults may have been unclearly specified when the lists were compiled and, in any case, must have been of dubious For the reasons mentioned carlicr, dita of this kind for cities would be dificult to interpret on account of the effects of migration, even if the records were fully complete and accurate, and it was known exactly how they had been compiled. As an illustration, Table 9 sunumarizes what is perhaps the most promising series of this Kind. Ie relates 9 Zirich in Switerhand which has for mane centuries been one of the most important commercial centres of Europe. It the seventeenth century lists of all che inhabitants were periodically drawn up bby the clergymen. Some ofthese were analysed inthe lst century by Daceynska. More recently medieval taxation lists which contain indications of marital taros hhave been studied by Schnyder. The tax levied in 1357 was « property tax only, ‘butin 1467 there was a head tax as well as a property tax. The head tax applied to ll persons aged 15 and over, 0 that in theory the figures for 1467 refer to 2 clearly defined age category. The data for 1637 are for the population aged 16 over. ‘The general piceure in Table 9 is very roughly constant through the cen 7 Sez Moy VoL I pp. FUROPEAN MARRIAGE PATTERNS JN PERSPECTIVE 07 Taste 9 Adult population of Zarich (Switzerland) by marital status Yor 1357 wer 17 ‘Number ra a8 ants Single } a ° Pee cened Wilowed ” 2 t Maced & o 2 Wane Namber 962 ap 2074 ‘Single “6 2 a er cened Widowed 5 7 3 Maret o “ a ‘Soure:Schayaee (520), pp. 56308 7 Decry (889), pp. 387208. turies, The proportions single among women are high; but there is every season to sospect a high rate of migration into the city which was a tiny enclave in an overwhelmingly rural population. (The age disribution in 1637 gives clear indication of immigration from the teens onwards.) The great excess of women over men suggests that the balance of immigration was particularly great for women. Large nuaibers who are single are thus no proof of abstention from marriage. The very small numbers of widowers and even edwin the Middle Ages (though notin the seventeenth century) are surprising in view the wey igh wan moray ofthe dine. I comely recorded hey Smply high rates of remarriage. But very high rates of remarriage for the Trek wey lng munbers ang ll sake sn old ebaunr Gond {tbe that the distinction between a widowed and a single person was not always recorded? The number of matred persons isin analyses of records of tis type deduced from the number of couples listed together; the distinction between the widowed and the single depends upon the recording of a fact which is not usually relevant to the administrative purposes of the listing, but which may emerge, for example, owing to the presence of children. ‘hed fr iclated meer eae ths oie ve for or pare a Similar difculkes apply with even greater force to various fragments avaiable forthe eaty Middle Ages at last the fornia which they are available so far in published work. A survey of these materials has been made by J.C. Russell (1938), who draws various conclusions from them which seem to me to be largely unwarranted. Z a ‘The most solid body of evidence concerning marriage frequency in Middle Ages is provided by the English poll tax records of 1377, analysed by Russell in his earlier book British Medieval Population (1948). The special tax cof one ‘groat” was levied in theory on all adults over 14 (except for the cleagy 4 Pechaps in some cates, some cs Ba J» this wend Wally sccacie gulag tad he pel ee he wer ers oe ltawn if tre were good information about martape in les Woald have tobe carly weighed, {i erage inate of the hrpe ates ote Ree camry mig bem irdon ook cera te marcage othe pion, but ofthe beginlngs of wew mariage babis which thea Spend Rosgheut Barope8 AgNaL who were separately taxed, and open beggars) and many of the tax lists have been preserved. It is possible to identify married couples in some of them. (There seems to have been no uniform pattem for drawing ap the lists of pessons who paid.) Russell analysed such lists for a number of villages and ‘towns. Hismain results are summarized in Table 10. In addition he has assembled data from parts of London boroughs and other places amounting in all toa few thousand further persons. The proporsions martied are similar to those in Table 10. Tasce 10 Percentages married among thse aged 14 or more from the English Poll Tax Returns of 1377 (villages and towns of varying size) ‘Sie Gi plae ge le Name of ‘Total perme Por cet of ted sot. of ‘aon taxed ‘pevon maried faba) ) Met” Women fas a = ws 230 to a 8 st-t00 1380 ™ oro 20 ” n sorqoo aan & & co Dacunouth 395 Ce) fer Cade ona 8 6 as ‘Kingstoa-oo-Hall () 2350 & s es Calthener() 3510 @ a Sours. C. Rael British Maeva Ppultion, Taba 7.277). ‘Nats: (@ Ths col nomber of istabitane is obained by Ruse by spplyng s ceasat icter of cases, (© The total popelations of Kington and Coleone wece calculated fiom 1,57 and apss enoas taxed, bot in ve ces sox and matial mars cold not be ceblsheds percentages macied ae bued ed 130 sad s.r person ‘The interpretation of these data is not subject to the dificultes arising out of small numbers or the ditonions due to migration. So far as coverage of towns and villages of varying size is concemed, we seem to have here better daa than any available for any country before the era of the modern censs. As has already been mentioned, the clerical population are omitted from the tax rerums summarized in Table 10. In an eatlier paper Russell (1044) made estimates of the clergy using Domesday, records of the specal poll taxes levied fon the clergy and other sources. In total the number of male clergy appears to hhave amounted to under $ per cent of the population aged 14 oF over. ‘There were some 2,000 nuns in a total of some 700,000 women over 14. Contrary to popula belief, the number of women taking the vell was very small; indeed nuns appear to have come mainly from the upper castes The ‘percentages married among men would thus have been a few per cent lower than those suggested by Table 10 if the clerical population were taken into secount. For women inchision of nuns would make no appreciable diference, 2 See ao Eilben Powe, Medial Engh Nunn EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE 19 ‘The percentages of married persons in Table 10 are much higher than could bbe expected on ‘European’ standards. It must be remembered that the widowed are included with che unmarried and that the vast bulk of the population was in the villages. Suppose we assume that the data are unbiased by marital status, i.e chat che chance that a non-clerical person of over 14 be included in the tax lists was the same for persons ofall marital conditions. Then the percentage of married among women over 14 in the population from which the poll tax samples of Table 10 are drawn could not have been under 67 per cent and was ‘probably somewhere meat 70 per cent, The per cent married atnong those over 4s must have been over 70 per cent (though knowledge of age at this period ‘would hardly have been very precise). On the European pattern, the Seon cet 15 who were married in a counery as ¢ whole ws below $5 and usually below so in the nineteenth century. ‘The percentage of women married according to the poll tax records is thus of quite the wrong order of magnitude for a population of European pattern ‘On the other hand itis very definitely of the sight size if the marriage pattern ‘was non-European, For example, in censuses taken in 1900 the following per- centages martied (among women aged 15 and over) were recorded: Bulga! — 6 Rumah «5 sata Unfortunately there seems no way of knowing whether the chance that a non-clerical person be included in the tax lists varied with marital sarus. On general grounds i is possible, and perhaps probable, thatthe paupers who were sctrbferinrarand hoe who wareael hoghkereohe ile sured wee lngelysnmarind. Asong hoe nw he lover gs i the sartied could pechaps more easly escape taxation by claiming to be too young. Twas perhaps sometimes posible to conceal the of unmarried relatives, while the head of che family and his wife could hardly be missed, Russell believes thatthe total number of paupers and untaxed is only of the order of 5 per cent, If half of them were women (in fact one is tempted to assume that fewer than balf were women) then the true percealage matried would stil have been very definitely ‘non-European’ even if all thote omitted were un rarried. Only if the number of omissions was very large with a heavy pre ponderence of the unmarried among those omitted could the populations ‘covered by the poll tax records of 1377 have had a European marziago pattern, If, on the other hand, the proportions of Table 10 are within a feww per ‘centage points of the truth one is tempted to conclude that the mariage pattern ‘of at least some parts of medieval England in the fourteenth century was not at all lke that ofthe cighteenth-century Europe, but much more like that of zon-Encopean civilizations. Further work on the poll tax records is needed before much confidence ‘could be placed in such an interpretation. Is there anything abnormal in those These genclinsione me bute oi dats Gm ceases of 850-9 elle ia Armate Inna Seis in), Vo" el popes (ar), Palka Petterage of women mace in the popeagon aged 14a over some pe cet below ‘the perectage marcel lathe populadoa aged 15 und overHAWAL few tax lists where married couples can be distinguished? Did che villages covered by these lists have any special character? The poll tax records are by far the most important data so far available on marriage patterns in the Middle Ages. Russell’s pioneee work on them needs to be extended with more detailed seudies based on the original documents. ‘Another type of evidence, the ‘inguisitions post mortem’, ic. legal docu- ‘ments involved in establishing claims to inheritance, has been used by Russell, and cited by others, to support the contention that women married relatively “ate in the Middle Ages atleast in the social cass covered by these documents. It's possible to clasify the inquisitions by age and by the marital satus of the heiress; and one can then attempt to determine the age at which one half of the heiresses are married. Statistically, chs procedure is closely analogous to the methods in the biological assay of, for example, insecticides. Groups of insects are exposed to treatment with an insecticide at various dosages. The dosage in biological asay corresponds to age in the inquistions. With inercasing dosage an increasing percentage of insects die, ust 2s with increasing age an increasing ‘peroentage of hiresscs arc martied. The problem is, inthe one case, to determine ‘what dosage is suficient to kill just 50 per cent of the insects, and inthe other, ‘at what age just 50 per cont of the heiresses are married. The statistical problems ‘of such data are tricky; and it might be interesting to attempt an analysis of the inguistions by the techniques of biological assay. Casual inspection suggests that che numbers Russell has been able to collect are far too small to support his conclusion, that in the reign of Edward Isome 50 per cent of eiresses were sil ‘unmarried at 24. Apart from the question of numbers, chore are several other ‘uncertainties in the interpretation of the inguistions (ebe accuracy of the determination of marital staus is suspect, etc). In any case, we now have the peerage materials zsembled by Hollingsworth (see Table 6) and the inquistions seem to fit in broadly with the pattern revealed, on much more solid evidence, by the peerage data. ‘We retum below to some indirect evidence relating to marriage in the Middle Ages. That some change in marriage habits took place between the fourteenth century and the eighteenth scems scarcely in doubt. In the Middle ‘Ages the betrothal of children and the marriage of very young adolescents ‘were apparently widespread throughout the population (not only among the nobility). ‘These practices had almost entirely disappeared by the eighteenth century. It does not seem possible chat the populations of medieval Europe had the fully developed European marriage pattem; they must either have had a marriage pattern clearly classifable as non-European, or else some mixture of the ewo types with a wider variation of age at first marriage than is found later. The ancient world As soon at they ae fourtsea, women az cilled “aie by men. ‘When they sce tat theis only resource is « be mariageable, they ‘begin to dll up and pu al thts hopes in that ‘Enctsrus: The Manual (and censry) Surprisingly, there is some statistical evidence on macriage in the Greco Roman world. One source of information is provided by the inscriptions on EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE. war tombstones. These often state the age at death; in the case of a mattied person dying the number of years he or she had been married is also sometimes stated. ‘By subtracting the duration of marriage from the age at death one can compute the age at marriage. Harkness in 1896 collected 171 inscriptions for women and 191 for men from which age at marriage could be calculated. The only ‘more recent attempt to assemble such data scems to be that of McDonell (1913) who covered.a far smaller number. With the enormous increase in the number of published inscriptions since that time, a far greater collection could now be obtained (see Russell (2958). Tapte rt Distribution of women's ages at marriage derived fom tombstones “Roman Inipone Cott Nonway ra No.of Percent Meese Gaus) Irons fe spine preen a » 7 = @ 3s a 2 36 is a ° 7 ” a 5 Pn ° 5 2 8 7 ‘Soe: Hates (18), Gasser snd Henry (950), p. ty and Norway, Satisike Centalburea, bel dnd flemaentons hag, 2066 ‘is, Norge Ofte Sei, Tree Ree, No, 106, rina 189, p43 Table 11 summarizes Harknes’ data for women. The sharp distinction which separates the distributions derived ftom the inscriptions from the ewo “European” distributions added for comparison is evident at a glance. The distributions derived from tombstones is certainly biased. The distributions of ages at death derived from inscriptions clearly do not represent the distribution of deaths in Roman times. For example, the deaths of young women are very probably over-represented (ee Durand (1960) and Henry (1957 and 1950)). A further bias is almost certainly introduced in the selection of the very small sp of all tombstones from which the age at marriage can be calculated. Probably the bias isin favour of those who mazried young; one reason may be that the duration of marriage is most likely to be recorded by the surviving spouse if the marriage has lasted 2 long time. (On the other hand matzied persons who dic before their spouses are likely to be old relative to their spouses) ‘When data of this type are published in the fature, a cross-tabulation of age at death by duration of marriage should be published and not only the distribution of the computed ages at marrage. ‘Whatever the bias, the conclusion to be drawn from Table 11 for our present argument canmot be in doubt. No conceivable bas in selection could produce from eighteenth- or nineteenth-century European marriage data the 2 The ean fr Cra i fr centages forthe higher age groves smanings only. if memarcinge were icine the (nnd oaicatiy =naa HAJNAL distributions yielded by the inscriptions. The population whose deaths the tombstones record had a marriage pattem of ‘non-European’ type. ‘The same conclusion emerges from an entirely different type of data, the ‘censuses’ of Roman Egypt. These were enumerations made for levying taxes. Declarations were required listing all the members of each household. Some ‘two hundred of these declarations are preserved and a full stady of hein has recently been published by Hombert and Préaux. The declarations give in- formation on relationship between members of each household and also data fn age. Something zbout age at marriage of women can be deduced by sub- tracing the age ofthe alr cd pres fom the age ef he mosh, he ‘Sgure so obtained will in general, be higher than che trae age at marriage and sometimes much higher (eg. when the first child had died or left home). Nevertheless Hombert and Préatx found (pp. 160-1) that of 155 women known to be married, at least sr had married before 20. This shows marriage at 2 ‘much younger age chan in the ‘European’ pattern, According to the data from Roman Egypt, husbands were usually con- siderably older than their wives. This suggests that men may have married ‘aicly late in life, That men marred late and indeed often deliberately abstained from marriage permanently both in Greece and later in Rome, has been ferred from literary evidence (notably Polybius) and from the Roman legislation designed to stimulate marriage. Ir has been disputed whether these trends ‘extended beyond the upper classes, Some fragments of statistical evidence were assembled by Landry (1936). Non-staistical evidence {have no fh in anything short of actual meatutement and the Rule of Three. ‘Cuuamuas Danwemi: Leter to W. D. Box (1855) ‘Was there, as the scanty statistical evidence suggests, a fundamental change jn marriage habits over much of Europe between 1400 and 1650? Ifo, where did the process begin and by what steps did it spread? What social or economic changes caused the new pattem of marriage to emerge? So much is known about the period, and indeed about the Middle Ages, that there ought 0 be plenty of materials for answering such questions—or so it would seem to one ‘who is not a historian, Ifthe characteistically late ‘European’ age at marriage should eum out to go back before the Middle Ages, the prospects for discovering its origins are, of course ess promising. It is probably too late now to determine whether Tacitus’ starement™ about the late age at sextal maturity among the The rer page (Gri son 20) sds as flows Gath Lob cit) ea jena venus cogae ec ving fesinarcur; ede oven sl procag: re is tempting for one who har to rly an a disionary aad Lats {ee Tge ob ele “Te lovely men ep oo be ee og a: “ose men began prog a yome sway, Not ate the gts shed (ao mariage), The Wein youth © OS nd song. The ples Dicioany EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE 133 Germans had any basis infact ot even exactly what he meant. It is hard to see hhow be could have had any sound information on such a topic. ‘The proposition that arguments about statistical matters from lit evidence ae rdky can be ilusrated ffom discussions of the age at mardage, ‘Thus Russell has pointed, in support of his view that people married late in the Middle Ages, to the doctrine thata man should have a living before he marvie, ‘Bursuch doctrines can be found in the literature of other civilizations. There is for example, a statement in the Talmnd that ‘a man should firs build a house, then planta vineyard and after that marry’. Yee it can hardly be doubted chat the Jews of Talmudic times married young. ate ier ri erry creas of cours he agen of tes “Late marriage’. In many societies with a distincely non-Earopean marriage pattem women marry late by Indian standards, but very early by European lf hey were looked for, promising lines of evidence which are not direcly scatsical might well eum up. An interesting possibility is suggested by the ‘work of Backman (1948). He briefly reviewed the references, mainly in legal literature, to the age of puberty, and concluded that the average age at menarche shad been constant since classical times at around 14 years, but that about 1500 1 process of retardation set in throughout Europe. It is posible that changing opinion ing the age at puberty and changing definitions of legal ajo afi 100 ceed phytologea changes Hamenee ico oto as likely also that we have here the accompaniment in legel thought of the social changes related to later marriage. On either interpretation the type of ‘material reviewed by Backman may well yield evidence chat a change in age at marriage occurred some time around che sixteath century. Clearly changes in the laws relating to marviage (including the minizasm legal age) are of interest in this connexion. A related matter is the controversy about the nature of mariage at the Reformation, Could the controversy have bbeen in one aspect an attempt to adapt theology to social change? Some of the themes (the voluntary agrecment of the spouses, parental consent) might clearly arise at a time of change towards later marriage. Ic is stiking that the properly organized registration of marriages spread during the sixteenth century over almost exactly the area where the European marriage pattern is later known to have prevailed °S ‘The interpretation of statistics in the matter of marriage has sometimes been distorted by general historical preconceptions, An interesting example is pro ded by temas (Doncdy ad Feel Sut, p. 10) who seid the 1379 polltax data for one sinall area (the ‘wapentake’ of Claro in Yorkshire). He found that the percentage married was much higher than in nineteenth-centary England. He concluded that the data must be at fault and the unmarried must. have ben especially prone to escape taxation. (He had other reasons for suspect- ing significant omisions from the tax lis) ‘Preconceptions drawn from the modera world have also influenced Homans? conti ni tuntncs os cea a ay Cates Ys 9 oem gen os ‘Babylonia Talnad, Tracate Souh 44a (queted in A. Cohen, Fomyman's Talal (London, ‘eat 3953) p37) ‘8 Seely, Vol pp. pda,ng HAJNAL work on English Villages in the Thirteenth Century which seems to be by far the most thorough study of marriage in a medieval rural population. Because the ‘rule was observed that a man could marry only after he acquired land, Homans infers that the age at marriage was high. ‘If 2 man had to wait undil his father died or gave up his holding, he would be likely to marry rather late in life’ (p. 158). Under conditions of high mortality such as existed in the Middle Ages, the conclusion does not follow. Over half the children in thirtcenth-century. England may have lost their father before they reached their 17th birthday. If they all married on their 17th bicthday and the remainder immediately after ‘their father’s death, there is no reason to think that the average age of all men at first marriage would have been above 24.% In fact, of course, as Homans documents in detail many were married during theit father’s life-time when the father turned over the land to them. Other aspects of Homans’ argument for a high age at marriage are volnerable on similar grounds; for example the analogy of moder Ireland” is misleading among other reasons because of its far lower death rates. When Homans wrote litte was known in detail about the mortality of pre-industrial Europe, or other underdeveloped societies. The ‘zequency with which in former days men’s lives were disrupted by the death of a relative is very hard to grasp for those living in modem Western societies.** Biicher’s litele monograph on the position of women in the Middle Ages —originally a lecture delivered in 1882—was expressly based on the belief that the problems of women, and in particular of single women, in his own day were similar to those of medieval times. All his evidence relates to a few cities ‘where indeed there was a substantial surplus of women. These cases are atypical not only because all towns were atypical in an overwhelmingly rural population, ‘but because Biicher’s data refer only to some of the largest commercial centres of extensive regions (Nuremberg, Frankfurt, Basle). There is reason to believe that the population of such centres was heavily recruited by immigration and that the female surplus then (as in more modern times) was larger in these cities than in smaller towns. Biicher ignores migration; he believes that the female surplus he found was due to excess male mortality which he attributes mainly to wars and debauchery. Eileen Power's essay on women in the volume on The Legacy of the Middle Ages (1926) draws on Biicher. It is impossible to say whether without his work. she would have made her categorical starement that ‘it must not be imagined that marriage was the lot of every woman and that the Middle Ages were not as familiar as our own day with the indey spinster’, She does, however, hint ar additional evidence not derived ftom Biicher, and, moreover, evidence sqace Tag oy brad arms an eb er on cer re ten te a ad ce ose. Sg mi Seoiveale a iulgecracay cctuccmms “eThe sriking calcaons en Alene Qos) tered te'sbors showing foe example how rl ea Ser tm zie try owing eagle hey ai mem ea oe moh ory nt ee Py EPL cena mover ci em mee oe cet Sabet ena ear eras tote EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE ws relating to the rural population: ‘A glance at any manorial “extent” will show ‘women villeins and cotters living upon their litle holdings and rendering the same services as men; some of these are widows, but many of them are obviously tuimarried.’ However, she does not present her material in detail; she seems t0 ‘expect the reader to feel that a high proportion of spinsters is not an unlikely state of affairs. ‘The interpretation of the history of European marriage should be informed by an awareness of the almost universal prevalence of other marriage patterns, and, preferably, by some knowledge of how they function. For example, Professor Ashton (English Economic History, p. 9) has argued that marriage rates rust have increased in the early eighteenth century because improved com- munication made it possible for people who would otherwise have been prevented from marrying to find partners. But most human societies through- out history have achieved nearly universal marrage at young ages with far worse conditions of ‘between human serdements than those of teven= teenth-oentury England. If the mechanisms for performing this feat no longer ‘existed in the seventeenth century, when did they break down? Was the cighteenth century, with a pattem of late marriage, preceded by a differenr pattern of even later marriage in the seventeenth ’—a pattern which departs even more radically from the norm of other societies? Ifsuch a remark- able situation existed in the seventeenth century, how did it originate? The hypothesis formulated by Professor Ashton is conceivable, bur it is not ren= dered plausible by his type of « prio’ argument. The age-sex composition ofthe population ‘When the baby i bom fits «boy, lt ves iis a gil expose it (od. From leer date +s, wrcen by Hiation to Als (presumably his wi}¢ ‘Though there is an enormous volume of literature on marriage and the family, lite of it has searted from the statistical end. Answers to some of the simpler questions raised by a statistical approach scem not to be available. Hlow has ir been possible, in most societies, to arrange for every woman to ‘marry? There ate several questions here; onc question would be answered by 2 comparative study of the various mechanisms by which marriage partners ‘were found. ‘The vast majority of human beings have always lived in small communities, such as villages of a few hundred people beeween which move- ment was difficult and often hazardous. The number of potential marriage parmers must have been stall and often diminished by rules (e.g. chose of caste) or conventions (eg. of clas) restricting the circle of those whom its proper to marry. How was it posible to find a partner for every gil? ‘The mechanisms by which this can be achieved must operate in very Uae fala esis te Stas Cay po aes Min ‘eich tnd in the oppste tron le thot ta there may hE sed toate erwecn 1377 au Thom whe fa the soteeth and seven ccary thee may have beens ‘oremeuttonadt eter mardage There wat a none ber Ofte lands pli ‘sho ad no incentive poane matte126 HAIWAL diferent ways in different societies. For ous purposes, the most relevant question is, what is dane in dificult cases where no suitable match fora girl secras avail- able? Comparative studies of the various devices by which this need is met would be of great interest. (The professional marriage brokers of Eastern European Jewish communities are an example, see Katz, (1950)). No doube a relevant factor which distinguishes modern Westem populations from the majority of societies is the convietion that marriage should be decided upon only after the furure spouses have got to know cach other well. This may render the finding of 2 mnacriage partner very diffcal since people often have opporaunities to become acquainted only with a few young persons of the opposite sex. If by contrat, i is posible to arrange a marriage between people ‘who have never met, the circle of potential spouses is greatly widened. It be- comes feasible to undertake lengthy journeys to trace a suitable partner, 2s when Abrabam sent his servane to travel to a distant country to find a wife for his son (see Genesis, Chapter 24). ‘The restrictions (e.g. of cxste), which lime the number of potential spouses for a given person in the village where he lives, operate also to designate the people among whom a parmer may be sought elsewhere. These restrictions, therefore, do not have the effect of keeping women unmartied; such restric~ tions do mean, however, that 2 network of kinship relationships is maintzined across larger areas than would be necessary if ll those living in one place could intermarry. They thas have an important effect both on the social organization and genecic structure of human populations A particular puzzle raised by those societies where almost 100 per cent of ‘women marry, i the fate of girls who are seriously handicapped by physical or mental disease or deformity. Though the line between light defects and those ‘which seriously affect a person’s chances of marriage is hard to draw, there are in modem Western societies probably one or two per cent in this condition (blind, deaf, spastic, epileptic, mentally defective, etc.). In other times and places, che scars left by malnutrition, disease and ignorant ereatment must have added to their numbers. On the other hand, many of those who were born with handicaps or acquired them when young probably died before reaching mar- iageable age. How were partners found for the rest? The answer again, mast ‘vary widely. In polygamous societies such women can become secondary wives. Elsewhere they may constitute the bulk of those who never marry. No ‘one seems to have discussed the problem. ‘We have so far discussed the methods by which spouses may be found for paticular people. How can the overall account be balanced, if there are surplus ‘women? In Europe, it has been customary to think ofa female surplus’ and the zesulting spinsters as a normal condition occasionally aggravated by war. Wars have not been unknown in other societies.” How did they find enough men to have every woman married? A tempting answer to this question is ‘poly- * ay eigen wh ct wt unieral manage of women foe grata de with rave stage of cn doe to war lowes. In a pe A cng des ob sald Ges of ts Scue wo his aed wit Sitar widows i tion rote nororionsXantippe. (Se Aled Zimmern, The Ge Com ‘nerve, Sh edion, Oxford, Chuendon Pres 193, Fe 340) | EUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE 27 ‘gammy’, But it is probably not an important factor in many societies which permit the practice; for it is often on too small a scale. ‘The ratio of the total number of men to the total number of womea in the population, though frequently mentioned in this context, is not very relevant. Nor is the ratio of adult men to adult women. For example in England at the ‘present time, there is an excess of over 1$ million women over 1 compared ‘with the aumber of men over 15. Yet so far as availabilty for marriage is com cemed, there is 2 shortage of women. The ‘surplus women’ are mainly widows over 69. ‘The first point to consider then isthe ratio of male to female population at the prime marriageable ages. It is probable that in cighteenth-century Europe this ratio was much les favourable to women’s chances of marriage than in many non-Western populations. There is, of course, always an excess of boys at birth, some 105 male births for every 100 female births. Male mortality is heavier than female mortality and in eightcenth-century Europe the excess of sales dying was probably sufficient, not only to produce equality of the sexes by the marriageable ages, but to create a female surplus. In the Scandinavian countries there was a considerable female surplus. By contrast in non-Western countries there seems nor infrequently to be a shortage of women. A famous ‘example is India. One theory (put forward, for example, by Coale and Hoover) is that women have been underenumerated at censuses to 2 greater degree than men. Other commentatots on the figures (e.g. Jain, 1954, p. 26) have, however, taken the view that the excess of men is gentine and can. be explained by sup- ‘posing that preferential treatment given to boys causes the death rate of boys to be lower than that of gicls. There have been similacly divergent interpreta- tions of the excess of men observed in several Chinese populat'vas. However, the excess of men in the population and the excess of female mortality is found aso in the data for Formosa at che beginning of the twentieth century and these data appear to have been of good qualty.*! Moreover, a good deal of data on death rates in underdeveloped countries has become available in recent decades and two world-wide comparative surveys of mortality (by Stolnitz and by the ‘United Nations Secretariat) have shown that where mortality is high, an excess of female mortality in childhood and at young adult ages is not infrequently found “We may, therefore, accept it as probable that in many non-Western popu lations there has been an excess of female mortality and a shortage of women at marriagcable ages. There are traces of the same tendencies in the data for Southern and Eastern Europe for the nineteenth century and beyond. The ‘impression of a considerable surplus of women in the population and a higher sale mortality seems to derive mainly from the statistics of Northem and ‘Western Europe. In the Middle Ages and eacier a shortage of women, or at anost a slight excess, was perhaps typical all over Europe. Laplace found that parents in the cighteenth century abandoned girls more often than. boys to 1 See pape by Tchr and Tosser (939) and Tar (960- {Th pcouenna i separate Som the queen of high taremal mortality, The es of ‘he aver hae Boe widely Sggented. :2B HASNAL the Foundling Hospital in Paris, The preference for boys may well have been at work on a larger scale in earlier times, ‘The ratio of men to women at matriageable ages does not, by itself, deter- ‘mine the numbers of men and women who can be martied in the course of their lives, For single women need not marry single men; they can marry widowers. If many marriages are dissolved by death while the surviving spouse is young, and if widowers remarry far more frequently than widows, it is posible for every woman to get married at lett once in her life even in a population where the number of men falls far short of the number of women, In this respect the remarriage of widowers works like polygamy. The point is not new, but often overlooked. Two hundred years ago, Sissmilch (Vol. I, p. 281) called the tendency to remarriage ‘polygamia successiva’. For him the issue was theological. The view that the excess of male mortality neatly can- celled the excess of boys born and resulted in an equality of the sexes by around age 20 had been an important doctrine of ‘physicotheology’. I illustrated not conly that everything in nature was harmoniously co-ordinated in accordance with a Divine plan, but also showed that the Christian principle of monogamy ‘was in accordance with the intentions of the Creator and superior to the heathea practice of polygamy. Siismilch discovered, however, that in the data available to him from several pars of Europe, the deaths of males under 20 exceeded those of femnales under 20 by a greater margin than the excess of boys among births, Le. chat by age 20 there were more women than men living. In this dilemma ‘polygamia successiva’ came to the rescue, Sistmilch could argue that there was nevertheless harmony in the Divine plan which included provision for monogamy, since the tendency for widowers to marty spinsters compen- sated for the surplus of women. ‘The degree of successive polygamy is variable and makes it possible to combine different sex compositions of the population with universal marriage. In eighteenth-century Europe, there were probably often between § an 10 pet ‘cent more women marrying for the first time than men. (See Table 12,) The degree of successive polygamy can be rapidly varied in response to changing circumstances. The history of Formosa described by Barclay provides a striking ‘Tapce 12 Effict of remarriage in compensating for imbalance in mabers of men and women eee eee ee eet EEE Pere ft weit Bee of women marin Cary nd date ones ee Pecan (Gomamp: 748479 % ‘Sweden: r7s0r800 z ” Francs (Crah): 1676170 & ea Fermon 1906, a % 510 % % 2920 a %s 2385 & 5 1563 %° = ‘Sowees: Stile, Vo I, p. 183; Gill, p 295 Guwier and Henry, p 83; Bacay, p. 235. BUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE. 129 example of the magnitude of this effect, At the beginning of the century the nuraber of men aged 10-49 was about 20 per cent greater than the number of ‘women in the same age group. This shortage of women was in part compen- sated for by female successive polygamy, ive. remarriage was more frequent ‘among women than men, As the sex ratio in the population became normalized and the proportion of remarriages wasreduced owing to a decline in widowhood and divorce, the situation came to resemble European conditions more closely. ‘The relationship between the numbers of each sex and their chances of marriage is also affected by the difference between the ages of brides and bride- ‘grooms at marriage. The point is most easily understood by a consideration of some hypothetical simplified examples. We first make three assumptions. () All brides are married at 20 t0 grooins aged 25. (i) Men marry in any case, but the proportion of women who marry is determined by the availability of men. (iii) The number of births is roughly constant from year to year. Under these conditions the proportion of women married will be reduced by the fact that the available number of men is diminished by death between 20 and 25. Suppose that the system were to be changed and gitls aged 20 were to be amarried to men aged 20 (everything clse remaining the same). Under these circumstances, the proportion of girls marrying is higher, because the men ‘who die between 20 and 25 are now available for marriage. If on the other hand. ‘men married at 30 (while women continue to marry at 20) the supply of bride grooms would be further reduced, compared with our initial assumption, by deaths between 25 and 30. i It is chas seen thet changes in the difference between the ages at marriage of men and of women aay aeons for variation in the balance between the numbers of each sex. The effect of differences in marriage age is even greater if we suppose that assumption (iii) above (namely that births are constant from. year to yeat) does not hold. Suppose that the population is increasing, ie. the number of bitths is increasing from year to year. If women aged 20 marry men aged 25 they marry men born five years before their own date of birth. But five years earlier fewer births were occurring. A large difference between the ages at marriage of men and women in a population of high mortality with an ‘increasing number of births tends greatly to reduce women’s chances of ‘marriage. * seal hf oop marie a he hf says cares (inh as Sng ep sy ens ols ate a ee Err Bape tar et pene of Soe ce een Gaon pn mao Wi peso latte sora caas oper tae of mabe pg he created Of Oar Or eee ela Een tp py a lay eee tpn ee Sr yates Date rem n Xel ealyoy n cy Yaseen demas voy ve tee ae Pets awa easy fo mops net be dns wan oS BES ne ae Scepter caste pcr cede Penman aus Sper al oa Se bmp ey ‘ey mre iis one nora gem cece eae ESE ete tapen a nape ae as ‘Bane Wess ey Or wh ma al vane Amn ees wl vey130 maywar From the beginning of the cighteenth century onwards a number of forces probably combined over much of north-western Europe to reduce the aval, ability of men for marriage when compared wich the number af women, ‘The decline in death rates reduced the number of widowers and hence the scope for successive polygamy. For example, in Sweden by 1901-10, only 10 percent of all marriages were contracted by widowers, as against 19 per cent in 1750-1800. ‘This eype of decline fiom about 20 to about 10 per ceat i probably typical of auch of Europe. The decline in death rates indirecy reduced wome's ances of marriage by increasing the rate of population growth, The effec of this ‘was explained in the previous paragraph, ‘Thirdly, emigration in the rincreenth ‘century was a predominantly male affic. Finally in many countries there was a decline in the marriage rates of men, as well as those of women, in the lanes half of the nineteenth century, a decline which has sometimes been regarded as amit response tothe feling that population growth needed vo be euaineh as part of the same set of | ‘which brought about the spread of birth ‘control. (France and Ireland have exceptional imarriage histories and some of this paragraph docs not apply to them.) Asaresale ofthese developments, the proportion of women never marrying rote to levels probably unprecedented in much of north westem Europe Be the end ofthe nineteenth century. The effect was temporarily reinforced by the First World War. Hovrever, since 1920 the situation has been completely tvans- formed and the ‘surplus of marriageable women’ which had corns tc bene, garded asa permanent condition has given place to a shortage of women, Conclusion ‘he main theme of this paper is not new. It is one of the main topics of ‘Malthus’ Essey and indeed implicit in its very structure (especially im the revised version of the second edition). Maks devoted Book I of his Essay to ‘the checks to population in the ls civilized pars of the world and ia past ies snd Book Ito ‘the diferent states of modem Europe’. In Europe e traces agin and again the workings of the ‘preventive checks of moral restraint’ which ‘mmplies ‘principally delay ofthe marriage union’ and he contrast the condition of Europe with that of the peoples described in Book 1. ‘Was Malthus sight in thinking chat late mariage in Europe resulted in Jower birth rates, and hence lower death rates, than obtained among non- Baropean population? Whatever the matare of the causal contectn, his notions about the level of birth and death rates gain some sippore fom sodetn rescarch. European bith rac, so fr as we ean tall, were acy over 38 before the spread of birth control; in underdeveloped counties, they are “almost always over 40 and offen over 45. So far as morality i concerned dhe contrat is less clear cus, but there sems no record in European experience Fae oe afro ene ra rma he et Smee agate Sere ss tire resem me a i SE tes Uae Tie etme Boose aersney eee RUROPEAN MARRIAGE PATTERNS IN FERSPECTIVE wr since the eighteenth century of conditions such as those in India o¢ Formosa in the initial decades of the ewenticth century. ‘The way in which a non-European marriage pattem goes with non ‘European birth and death rates may be illustrated by a recent study by Csoscén of the parish registers for three Hungarian villages in the eighteenth century. This population isnot in ‘Europe’ as defined for this paper. The distribution of smarrlages by age of bride in 1770-1800 was quite definitely ‘non-European’ a8 the following figures show: Per at batt Age gop Brilgeon Bre Under 30 " 2 2m “ 25 30 1% 8 m9, 8 3 and over 2 3 Teal 100 so Tn the same period (1770-1800) the crude birth rate in these Hungarian villages was $2 per 1,000 and the death rate was 43 per 1,000. This may be computed, for example, with the following figures for the French village of CCrulai given in the study by Gautier and Henry: Porst Bik re Dest te rasp 36 7309 Py 8 ‘The marriage data for Crulai which have been mentioned several times ate, of course, quite clearly ‘European’. ‘The cighteenth-century Hungarian ‘villages are thus non-European in all tree respects (age at mariage, birth rate, death rate) in contrast with the European levels of Crolai’s vital rate, as well geese rica pene eee ene een re European conditions were fimdamentally different not oaly in marriage, birth and death rates, but above all in standards of living, from those obuining elewhere in the world. Europeans, a large proportion of them, not just the rich, bad better housing, better clothing, a greater variety of food, more furniture and utensils, chan people elewhere. This uniqueness of Europe, so evident to contemporaries, has been largely ignored in recent discussions of economic development; all chat is presindustrial, including cighteenth-century rope, often Ianpal togeher in geerzaons about ‘gular!’ or ‘peasant’ or ‘underdeveloped! societies. Presumably the uniqueness of Europe in standards of living and in death rates did not extend back beyond the seventeenth century (except in limited regions), But if European death rates were as high asin other parts of the world, ‘ould birth rates have been lower? And if the European ith re bere he seventeenth century was as high as elsewhere, does this not imply that European Women must have married young as in other populations of high fertility? 4 These are based on 44o cates for men and 442 for women. "The namber of mar- sing ress bono os a fe wo Then ae Seg ES yeaah see ote pr oe eee sear| : | ne HAyNAL ‘These large and vague questions need to be broken up and investigated by ‘atefol calculations on the interclationships beeween marriage, birth and death rates. Even the relation berween marriage patterns and crude birth rates is not neatly 28 obvious as is often supposed and it is not independent of mortaliey. An inquiry into the origins of the European marriage pattem will in cvitably take one into fundamental isues of economic and socal history. This is so not only because of the connexions just discussed between marriages and Births and deaths. There are other links." marriage almost by definition te- quires the establishment ofan economic bass forthe life ofthe couple and theit children. The arrangements curtent in a society for achieving this must fit in with the marriage pattern: they will shape it and will bein turn influenced by 3 Unmarried men and women must be attached v9 houscholds in some way, or form independent households. The structure and size of households and the tate of formation of new households and disappearance of old anes, therefore, depend on the marriage patzer. Tn societies where the household isthe principal uit of economic production as well as consumption, all cis means that the marriage pater sted in very intimately with the performance of the cconomy. asa whole. The emotional content of martiage, the relation between the couple vsually a girl of 16 and one in which she is typically a woman of 24. These things are perhaps obvious, but they have not been much explored, at least nor in histories which trace the emergence of modem Europe, A full explanation of the background of European marriage patems would probably lead into such topics s the rise of capitalism and the protestant ethic, ‘The economic system influences the pattern through the sy wich 2 co tn eee ph age explored. In the European pattem a person would usually have some years of adult life before marriage; for women especially this period would be much larger than outside Europe It isa period of maximum prodactive capacity Without responsibility for children; a period during which saving would be cary. These savings (e.g. by means of the cumulation of houschold goods in reparation of marriage) might add substantially ea the demand for goods other than the food ccc. required for immediate survival. in this respect delayed ‘marriage may be similar co income inequality in stimulating the diversion of resources to ends other than those of minimum subsistence; but when later marriage is the norm the total volume of demand generated might be much larger than that which can be caused by a small cas of wealthy famalics in a Population at subsistence level? Could this effet, which was uniquely Buro- pean, help to explain how the ground-work was lid forthe uniquely Bucopean “take-off into modem economic groweh? j/ Top mete recace inthe hbo force of lge uber of adult women oot icwared ia of easng ms have Ben consdcrabe advantage to the eghasatcaners Bees ran economic, BUROPEAN MARRIAGE PATTERNS IN PERSPECTIVE 133 late marriage brings about wealth, wealth may equally cause late marriage. Te was suggested in che eighteenth century (for example by Cantillon) that people married late because they insisted on a certain standard of living (a Standard varying with the socal postion of the individual) asa prerequisite of ‘marriage, More simply, men marry late because they cannot ‘afford’ to marry ‘young; they have vo wait until they have a livelihood, a farmer till he aequizes land, an apprentice till he finishes his apprenticeship and so oa. It is tempting to seein this feature a key to the uniqueness of the European ‘marriage pattern. In Europe it has been nevesary for a man to defer marriage tui he could establish an independent livelihood adequate to support a family; jn other soccties the young couple could be incorporated in a larger economic unit, such as a joint family.* This, presumably, is more easily achieved and docs not require such a long postponement of maztiage. This lin of argument seems specially convincing i the larger economic unit is such that extra labour it often felt to be an economic asset. A system of large estates with large house~ holds as in Eastern Europe might thus be conducive to a non-European marriage ‘pattern, while small holdings occupied by a single family and pased on t0 2 Single heir would result in a European pattern. If this reasoning has subseance, the uniquenes of the European marriage pattern must be asctibed to the European ‘stem-femily’ (The term ‘stem-family" seine by Le lay a describing the type of family organization in which land descends to a sn heir, sete pg dy Trina cnc which may have helped to bring about the European marriage pattern, if it did noc ex in he Mle age ime tal eee al nee presumably a delay in the death of the holders of land resulting from declining death rates would tend to raise the age at marriage. Whether there was, in face, 4 decline in mortality over the relevant period is a dubious point (the decline jn question must have occurred before the eighteenth century); but this is ‘erainly a hypothesis chat merits study. ‘The connexion between the Sth 2 th ole fn ands avabey for the founding of a new family i, however, rather an indirect one. Under the mortality conditions of the Middle Ages fathers often died while their children were very young; interim arrangements had to be made till the son ‘was old enough to take over. Even ifthe father survived to eld age, ie does not The yung Da ee Rain woe eso ie i apd wee eed ie rans i ean apd ee ua GSA Ee ety ce ce Tao cane a Sha prc hy mh a ooh na ee pe ne a HED! fo which generation con jusiGably be made about the fay eaten fn pars of ep a Pa ae eh Wy a ch he ey pte nee aaah Sy re neat i ce ol gor crn ai) Ciclo ieee de fos Sr ceo an eee et sac aL i a sd tea sea oo ua al tenet prea sgt ual ost fence ety 9) i nen anata ee ‘Kaen ofa ttl Lb aco aged whale sees Oy Houscholds coming of several acer fires fey no retin ng i die canon tae ‘HEamgemen uke if pousle for young couples co be part Of 3 lng unto te begdacag of Iariage for some yes.wana follow that a young family could not be set up on the holding until he died. Homans in his book on thi England describes many instances where a father made over the land to his son while he lived, chus permitting the later to get martied. He also mentions instances where a father, while he lived, turned over his holding ro be shared berween two sons, where a man transferred his holding to someone other than his son, et. To understand the effect on the frequency of marriage and age of marriage of a rule that a man sist acquire land before marrying we should have to know the frequency of the various arrangements by which land was passed on. The rate at which land became available for the founding of new families may have been controlled not so much by death as by social arrangements Is nota all clear «priori how ‘arule that 2 man must have a livelihood before marrying would operate to ‘produce just such a postponement as is in fact observed, Even if we understood hhow the age at marriage of men was determined at a given period it would still need to be explained how women's age at marriage was effected. The uniqueness of the European patter lies primarily in the high age at marriage of women (often with a relatively small difference berween the age of husband and. wife), rather than in a high age at marriage for men, ‘There is no space for farther specalation on the causes or consequences of the Enropean marriage pattern, The primary concer of this account has been the mere existence of the pattem. This aspect should be kept distinct from the search for explanations. It has been shown (1) thatthe distinctively European patter can be traced back with fair confidence as far asthe seventeenth century in the general population; (2) that its origins lie somewhere about the sixteenth ‘century in several of the special upper class groups available for study and in none of these groups was the pattern European before the sixteenth century; (6) the lite fragmentary evidence which exists for the Middle Ages suggests ‘a non-European pattern, as do scraps of information for the ancient world. Some at leas of the data presented have probably been misinterpreted. In dealing with sources of a type of which one has no experience coming from remote periods of whose historical background one is ignorant, one is very likely to make mistakes. In an effort to survey so great a variety of materials, some of them could only be looked at superficially. Even if individual pieces of information have been soundly interpreted, there remains the problem how far generalizations can properly be based on isolated demographic facts. This isa basic problem of much of historical demography. We wish to draw con~

You might also like

• Fear: Trump in the White House

  

  From Everand

  Fear: Trump in the White House

  Bob Woodward

  Rating: 3.5 out of 5 stars

  3.5/5 (738)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  From Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Rating: 4.5 out of 5 stars

  4.5/5 (122)
• A Man Called Ove: A Novel

  

  From Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Rating: 4.5 out of 5 stars

  4.5/5 (4610)
• • Magazines
  • Podcasts
  • Sheet music
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  From Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Rating: 3.5 out of 5 stars

  3.5/5 (231)
• Grit: The Power of Passion and Perseverance

  

  From Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Rating: 4 out of 5 stars

  4/5 (590)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  From Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Rating: 4.5 out of 5 stars

  4.5/5 (842)
• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  From Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Rating: 4 out of 5 stars

  4/5 (5807)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  From Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Rating: 4.5 out of 5 stars

  4.5/5 (266)
• John Adams

  

  From Everand

  John Adams

  David McCullough

  Rating: 4.5 out of 5 stars

  4.5/5 (2409)
• Shoe Dog: A Memoir by the Creator of Nike

  

  From Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Rating: 4.5 out of 5 stars

  4.5/5 (537)
• Rise of ISIS: A Threat We Can't Ignore

  

  From Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Rating: 3.5 out of 5 stars

  3.5/5 (137)
• Principles: Life and Work

  

  From Everand

  Principles: Life and Work

  Ray Dalio

  Rating: 4 out of 5 stars

  4/5 (608)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  From Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Rating: 3.5 out of 5 stars

  3.5/5 (401)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  From Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Rating: 4.5 out of 5 stars

  4.5/5 (234)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  From Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Rating: 3.5 out of 5 stars

  3.5/5 (2259)
• The Glass Castle: A Memoir

  

  From Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Rating: 4.5 out of 5 stars

  4.5/5 (1716)
• Yes Please

  

  From Everand

  Yes Please

  Amy Poehler

  Rating: 4 out of 5 stars

  4/5 (1898)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  From Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Rating: 4.5 out of 5 stars

  4.5/5 (346)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  From Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Rating: 4 out of 5 stars

  4/5 (897)
• Wolf Hall: A Novel

  

  From Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Rating: 4 out of 5 stars

  4/5 (3811)
• The Emperor of All Maladies: A Biography of Cancer

  

  From Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Rating: 4.5 out of 5 stars

  4.5/5 (271)
• Angela's Ashes: A Memoir

  

  From Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Rating: 4.5 out of 5 stars

  4.5/5 (440)
• Her Body and Other Parties: Stories

  

  From Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Rating: 4 out of 5 stars

  4/5 (821)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  From Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brene Brown

  Rating: 4 out of 5 stars

  4/5 (1091)
• A Tree Grows in Brooklyn

  

  From Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Rating: 4.5 out of 5 stars

  4.5/5 (1929)
• The Art of Racing in the Rain: A Novel

  

  From Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Rating: 4 out of 5 stars

  4/5 (4203)
• The Light Between Oceans: A Novel

  

  From Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Rating: 4.5 out of 5 stars

  4.5/5 (789)
• The Perks of Being a Wallflower

  

  From Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Rating: 4.5 out of 5 stars

  4.5/5 (2104)
• The Constant Gardener: A Novel

  

  From Everand

  The Constant Gardener: A Novel

  John le Carré

  Rating: 3.5 out of 5 stars

  3.5/5 (104)
• The Outsider: A Novel

  

  From Everand

  The Outsider: A Novel

  Stephen King

  Rating: 4 out of 5 stars

  4/5 (1850)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  From Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Rating: 4.5 out of 5 stars

  4.5/5 (474)
• On Fire: The (Burning) Case for a Green New Deal

  

  From Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Rating: 4 out of 5 stars

  4/5 (74)
• Sing, Unburied, Sing: A Novel

  

  From Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Rating: 4 out of 5 stars

  4/5 (1104)
• The Woman in Cabin 10

  

  From Everand

  The Woman in Cabin 10

  Ruth Ware

  Rating: 3.5 out of 5 stars

  3.5/5 (2521)
• Brooklyn: A Novel

  

  From Everand

  Brooklyn: A Novel

  Colm Toibin

  Rating: 3.5 out of 5 stars

  3.5/5 (1946)
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  From Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Rating: 4 out of 5 stars

  4/5 (98)
• Little Women

  

  From Everand

  Little Women

  Louisa May Alcott

  Rating: 4 out of 5 stars

  4/5 (104)
• The Unwinding: An Inner History of the New America

  

  From Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Rating: 4 out of 5 stars

  4/5 (45)
• Steve Jobs

  

  From Everand

  Steve Jobs

  Walter Isaacson

  Rating: 4.5 out of 5 stars

  4.5/5 (806)
• Manhattan Beach: A Novel

  

  From Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Rating: 3.5 out of 5 stars

  3.5/5 (792)
• Bad Feminist: Essays

  

  From Everand

  Bad Feminist: Essays

  Roxane Gay

  Rating: 4 out of 5 stars

  4/5 (1016)