UNIVERSITY OF ILLINO ^ IBRARY UP- '•■■■ AMPA' ■ r> ~ ^ A n • UNIVERSITY OF ^ ILLINOIS LIBRARY ' URBANA-CHAMPA' I i imt Digitized by the Internet Archive in 2011 with funding from University of Illinois Urbana-Champaign http://www.archive.org/details/testingofhousety11jakl I ^T n no ■ II t,lo OCCASIONAL PUBLICATIONS OF THE DEPARTMENT OF GEOGRAPHY THE TESTING OF A HOUSE TYPING SYSTEM rN TWO MIDDLE WESTERN COUNTIES: A COMPARATIVE ANALYSIS OF RURAL HOUSES by JOHN A. JAKLE PAPER NUMBER AUGUST 1976 SUSAN R. GROSS, editor GEOGRAPHY GRADUATE STUDENT ASSOCIATION UNIVERSITY OF ILLINOIS at URBANA - CHAMPAIGN The person charging this material is re- sponsible for its return to the library from which it was \vithdrawn on or before the Latest Date stamped below. Theft, mutilation, and underlining of books are reasons for disciplinary action and may result in dismissal from the University. UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN AUG 2 3 The 1979 is tngarten sen •erIn-Mandlebaum preparing the graphics, L161 — O-1096 I A COMPAEATIVE ANALYSIS OF RURAL HOUSES IN TWO MIDDLE WESTERN COUNTIES: THE TESTING OF A HOUSE TYPING SYSTEM* John A. Jakle** ABSTRACT A prototype house classification system was tested in an analysis of rural houses in two Middle Western counties. The house typing procedure first classified houses by structure (outline of floor plan, height as defined by the number of stories, and roof type) and then by architectural style or decoration. The most prevalent structural types were narrow and wide oblong houses of either one or one and one-half stories with end-gabled roofs. These houses, which I call I-cottages and double-I cottages, probably represent the mainstream of American house architecture. The vast majority of the houses of both study areas had no architectural decoration. When structural and architectural style variables were combined in an analysis of pre-1920 houses, only seven numerically significant house types were identified. Students of house types in emphasizing traditional folk houses may be focusing on an exceptionally small proportion of the total houses available for study. Development of a universally applicable house typing system for the study of entire house populations seems desirable. INTRODUCTION Houses usually differ substantially from one society to another. Within societies architectural emphases usually vary through time. This diversity reflects the wide variety of functions which houses are made to serve and the wide spectrum of social meanings which they convey. Houses not only shelter biological man, but they also shelter the social being, conveying messages about life style and social status. Houses are often *The author wishes to thank the National Science Foundation for supporting the geography field camp conducted in Parke County, Indiana, by the University of Illinois during the summer of 1970 and the Depart- ment of Conservation of the State of Illinois for supporting the 1974 field course taught from Urbana. I extend thanks to the many students and to Fred Foster, my faculty colleague, for their help in conducting the two house surveys reported herein. Special thanks to Max Michels, Tony Lazewski, and David Adams for computer assistance and to Jim Bier for the cartography and graphics. **Dr. Jakle is an Associate Professor in the Department of Geography, University of Illinois at Urbana-Champaign . 1975-76 Staff Senior editor: Junior editor: Business managers: Susan Gross James Altengarten Ricliard Olsen Randi Halperin-Mandlebaum The author wishes to thanic Jim Bier for preparing the graphics. A COMPARATIVE ANALYSIS OF RURAL HOUSES IN TWO MIDDLE WESTERN COUNTIES: THE TESTING OF A HOUSE TYPING SYSTEM* John A. Jakle** ABSTRACT A prototype house classification system was tested in an analysis of rural houses in two Middle Western counties. The house typing procedure first classified houses by structure (outline of floor plan, height as defined by the number of stories, and roof type) and then by architectural style or decoration. The most prevalent structural types were narrow and wide oblong houses of either one or one and one-half stories with end-gabled roofs. These houses, which I call I-cottages and double-I cottages, probably represent the mainstream of American house architecture. The vast majority of the houses of both study areas had no architectural decoration. When structural and architectural style variables were combined in an analysis of pre-1920 houses, only seven numerically significant house types were identified. Students of house types in emphasizing traditional folk houses may be focusing on an exceptionally small proportion of the total houses available for study. Development of a universally applicable house typing system for the study of entire house populations seems desirable. INTRODUCTION Houses usually differ substantially from one society to another. Within societies architectural emphases usually vary through time. This diversity reflects the wide variety of functions which houses are made to serve and the wide spectrum of social meanings which they convey. Houses not only shelter biological man, but they also shelter the social being, conveying messages about life style and social status. Houses are often *The author wishes to thank the National Science Foundation for supporting the geography field camp conducted in Parke County, Indiana, by the University of Illinois during the summer of 1970 and the Depart- ment of Conservation of the State of Illinois for supporting the 1974 field course taught from Urbana. I extend thanks to the many students and to Fred Foster, my faculty colleague, for their help in conducting the two house surveys reported herein. Special thanks to Max Michels, Tony Lazewski, and David Adams for computer assistance and to Jim Bier for the cartography and graphics. **Dr. Jakle is an Associate Professor in the Department of Geography, University of Illinois at Urbana-Champaign. the most personalized features of a settlement landscape and, accordingly, are often the most difficult of settlement features to classify. To date, attempts to classify American houses in academic geography have tended toward two extremes. Most studies have described the distribution of subjectively defined houses over broad geographical areas. Perhaps, the works of Fred Kniffen, Henry Glassie, Peirce Lewis, and Pillsbury and Kardos are best known. Each author has emphasized historically significant folk house types, traced their origins, and described their diffusion in the American scene. A few geographers, like Ken Corey, have used quantitative techniques, such as principal components 2 factor analysis, to more objectively classify houses. For logistical reasons their studies have tended to treat relatively small areas; for example, Corey analyzed only 250 houses in a single Cincinnati neighbor- hood. Most quantitative analyses have been limited to the use of variables obtained from secondary sources such as tax records . Great diversity of topical and areal coverage exists in the house type literature, yet no objective system universally applicable to the 3 description of houses has ever been attempted. John Rickert's check- list for identifying house facades by period of origin may prove applicable throughout the United States, but it was not intended as a means of systematically relating house characteristics for purposes of 4 house type definition. In this article, therefore, I offer an example of what a universally applicable house typing scheme might involve. A prototype classification was tested in the study of rural houses in two Middle Western areas: Parke County, Indiana, and Vermilion County, Illinois, Houses were first classified by structural type and then by architectural style or decoration. Houses built before 1920 were further analyzed for historically significant "house types" by combining the attributes of structure and decoration. I sought an objective system of house typing capable of identifying the traditional house types discussed in the literature. THE STUDY AREAS The house classification scheme was tested in two very different areas: one where a large number of traditional folk house types were expected (Parke County, Indiana) and one where they were not (Vermilion County, Illinois) . The two counties are located in the Wabash Valley on or near the boundary separating Indiana and Illinois (Fig. 1). The agricultural development of both areas began about 1820 with settlers coming to Parke County largely from the Upper South and to Vermilion County from both the South and the Northeast. Both counties grew at the same rate until the 1850 's when Vermilion County's population began to rapidly increase with the growth of Danville as a railroad center, the drainage of the wet prairies and their conversion to culti- vatioi\ and the discovery of coal (Fig. 2). Twentieth century "cornbelt" agriculture (marked by recurrent market shifts, resultant crop reorienta- tions and the introduction of new production technologies) has brought widespread and continual renewal of the Illinois area's rural landscape, Parke County's population, on the other hand, stabilized in the late nineteenth century and slowly declined in the twentieth century (Fig. 2). The area has never experienced large-scale urbanization and, given the preponderance of hilly terrain and thin soils, its agriculture has never Hoopeston VERMILION COUNTY Urbana Champaign Crawfordsville. 10 20 1 I I I ==i Lafayette 3 D T=^ PARKE COUNTY Greencastle Paris, Brazil Terre Haute Fig. .1: The Vermilion County, Illinois, and Parke County, Indiana, Study Areas. Houses were inventoried within the shaded areas on the map. experienced marked prosperity. Conversion of agricultural land to recreational use has been widespread. Whereas the Vermilion County landscape strikes the visitor as being very "contemporary," the Parke County landscape appears antiquated : a virtual museum of house architec- ture. population lUO.OOO y — 80 000 / — ^ ^- / r- 60,000 / Vermilion County, 1 Illinois / 40 000 / \ / 20,000 7 Par keO unty Indiana ^ yf' ^ ■ — — . j -^ 1 1 1830 40 50 60 70 80 90 1900 10 20 30 40 50 60 70 Fig, 2; Population Change in the Two Study areas, 1830 to 1970. THE HOUSE SURVEY Urbanized areas within the two study areas were eliminated from the surveys in order to focus on rural houses. Three state parks and the Mansfield Reservoir were also omitted from the Indiana study area and the strip-^ined areas south and west of Danville were omitted from the Illinois area (Fig. 1). Within those areas surveyed, all farm and rural nonfarm dwellings visible from public roads were classified; 2,357 houses in Parke County and 2,265 houses in Vermilion County were surveyed during June 1970 and June 1974, respectively. Survey teams composed of either two or three persons, were instructed to debate and reach consensus regarding house characteristics observed. CLASSIFICATION OF HOUSE STRUCTURE The house classification system used was based on a number of assumptions concerning how Americans view house exteriors. It was assumed that they conceptualize houses in two highly related yet very different ways. First, they see houses as having structural properties such as size, shape, and layout. One house may be big and another small. One house may look like a square box whereas another appears to be a multitude of geometric forms. Second, they see houses as having archi- tectural decoration above and beyond the basic elements of form. One house may be of Italianate design with bracketing under the eaves and arches over the windows whereas another may be of classical revival style with pillars and a large pediment. People may not know the appro- priate name for a specific style, but they see the decorative trim and differentiate it from a building's structure. In the past, architectural historians have tended to treat struc- tural form as a part of architectural decoration whereas geographers in focusing on structure, have tended to de-emphasize style. Many scholars consider structure and decoration as integrally linked. They observe that certain structural and decorative features are always highly correlated. I believe that form and decoration should be studied separately if for no other reason than to more clearly identify the circumstances where form and decorative styles "merge." In this analysis, house structure was classified first and decorative styling second. Again, house structure was seen to involve elements of height, shape, and roof line; architectural styling was seen to comprise exterior decoration attached to or covering a basic house structure. Another assumption concerned the significance of front entrances, I believe that in conceptualizing houses Americans orient themselves to the fronts of houses. Ask children to draw pictures of houses and they Q invariably draw front facades. I further assumed that where multiple house units are combined to form a single house structure, the section which contains the front door is seen as the main or principle house unit. The front entrance usually gives access to rooms used for the more formal entertaining of friends and relatives. It is here that the family's image is most deliberately manipulated in terms of interior decoration and this portion of the house, particularly as viewed from the exterior, is the family's principal symbolic picture to the world. Only the "main unit" of each house was classified. Appendages or extensions off main house units were not included. The latter were defined as separate house sections set off by different roof forms, dif- ferent height characteristics, or both. They need not have been built as additions, but could have been original to the house. Focus was placed on "main units" as they stood unadorned by structural elaborations, I did not intend to diminish the significance of house extensions for they certainly form a significant house characteristic which must be treated carefully in future analysis. For this study, however, it was decided to deal first with the very basics of house appearance and thus the simplifying strategy of classifying only main house units where multiple unit houses were found. Unfortunately, several house types, such as the upright and wing house (often called the temple and wing house when combined with classical revival architecture) , could not be properly 9 identified in this analysis. Nonetheless, only five and eleven per cent 8 of the houses in Parke and Vermilion Counties, respectively, could not be treated using my simplifying scheme (Table 1) . TABLE 1. FLOOR SHAPE OF MAIN UNIT Parke County Vermilio Number n County Number Per Cent Per Cent Shape of Houses of Total of Houses of Total Narrow Oblong (I) 1,003 43 493 22 Wide Oblong (II) 498 21 673 30 Square (III) 380 16 281 12 Irregular Massing (IV) 359 15 562 25 Polygonal or Round (V) 7 Other 117 5 249 11 2,357 100 2,265 100 SOURCE: Field Surveys. General Shape For the purposes of this study, the "structure" of a house was taken to comprise three basic dimensions: (1) general shape as defined by the perimeter outline of the floor plan, (2) height as defined by the number of occupied or finished floors or stories, and (3) roof type. The categories used to describe the first dimension, general floor shape, are diagrammed in Figure 3. The "narrow oblong" (I), whether of one story or of two stories, usually describes houses one room deep and two rooms wide excluding hallways (Figs. 3a and 3b), The "wide oblong" (II), in both the one- story and two-story examples , usually describes houses one room deep and two rooms wide excluding hallways (Figs. 3c and 3d). The "square" (III) in both the one-story and two-story varieties contains either two or three rooms along each dimension (Figs. 3e and 3f). Two-story houses with square floor perimeters tend to be cubic in shape. Irregular mas- sing (IV) is something of a miscellaneous category. All structures featuring irregular floor plans not better described in another category were considered here. Examples are diagrammed (Figs. 3g and 3h) . This category usually involves asymmetrical houses which combine masses or blocks of different sizes in informal ways. Irregular massing is not to be confused with narrow oblong, wide oblong, or square houses with added architectural adornments such as pavilions and towers. Nor should houses with structurally distinct extensions be confused with irregular massing. Irregular massing involves the use of different geometric forms in a single, fully integrated structural unit. Integration derives from the use of a single roof and the retention of a single height throughout. All houses in which geometric symmetry in the floor plan produced match- ing sides (without recourse to ninety degree angles) or where a single circular exterior wall was used were to be categorized as polygonal or round (V). No such houses were found in either study area. The numbers of rural houses found in each category in the two counties are given in Table 1. Of particular significance, the narrow oblong shape appeared nearly twice as frequently in Parke County (forty- three per cent) as in Vermilion County (twenty-two per cent) , This narrow oblong shape is found in the "I-house" first identified by Kniffen, As such, it was characteristic of the early nineteenth 10 NARROW OBLONG (I) / ^ / K(a) / Y A\i) WIDE OBLONG (II) / A V SQUARE (III) IRREGULAR MASSING (IV) (0 Fig, 3: Floor Shape Categories. century in a broad area stretching from Pennsylvania through the Upper South and Lower Middle West. On the other hand, the wide oblong (twenty- one and thirty per cent of the Parke and Vermilion County houses, 11 respectively) was more typical of the Middle Atlantic States and New 12 England during this period. Irregular massing (fifteen and twenty- five per cent of the Parke and Vermilion County houses) was universally typical of the late nineteenth and early twentieth centuries. Perhaps indicative of a structurally more complex house population in the more contemporary Illinois landscape, twice as many houses in Vermilion County (eleven per cent as opposed to only five per cent in Parke County) could not be classified in terms of floor shape. Height of House The vertical dimension was measured according to the number of occupied or finished floors: 1, 1-1/2, 2, 2-1/2, and 3 stories, res- pectively. A "half-story" was defined as living space created under the roof as indicated by window openings in dormers, end gables, or gambrels. The number of houses in each height category is given for both study areas in Table 2. Parke County houses tended to be smaller as judged by the number of floors. Only eleven per cent of the houses in the Indiana study area were two stories or higher whereas forty-two per cent of the houses in the Illinois study area were so classified. Roof Type The roof-type categories used are diagrammed in Figure 4. Perhaps the most common in the Middle West, they by no means exhaust the variety of roof forms found. Nonetheless, they were adequate to the description of ninety-seven per cent of the houses surveyed in Parke County and eighty-eight per cent of those in Vermilion County. Frequency counts for the two study areas are given in Table 3. The end-gabled house, more 12 TABLE 2. NUMBER OF STORIES ON MAIN UNIT Parke County Vermilio Number n County Number Per Cent Per Cent Number of Stories of Houses of Total of Houses of Total One 1,333 56 896 AO One and one-half 784 33 636 28 Two 207 9 506 22 Two and one-half 13 1 135 6 Three 1 Three and one-half 1 Basement only 1 2 Other 19 1 88 4 2,357 100 2,265 100 SOURCE: Field Surveys. common in the South than in the Northeast in the nineteenth century, accounted for fifty-one per cent of the houses in Parke County, but only thirty-four per cent in Vermilion County. The front-gabled house, more common to the Northeast, comprised seventeen per cent of the Vermilion County houses, but only ten per cent of those in Parke County, The substantially smaller number of houses unclassified in Parke County (three per cent as opposed to twelve per cent for Vermilion County) is also suggestive of a more complex house population in the Illinois area. 13 .'^ \ (0 Front Gable • /I ; : \/ (e) End Gable 3^ I _J__l J- ^^ T— (fc) Front I Centered Gable I I I I L-L I I I J (m) Multiple Gables L_--> \ /I v, /^ 1 r^ 1 1 • 1 / 1 / .....i' ' (g) Gambre (fh) Flat Hip ^^ (pr) Peaked or Ridged Hip (m) Mansard Figure 4: Roof Type Categories. Shape, Height, Roof Combinations Floor shape, height, and roof type characteristics were combined in order to identify general structural types. The twenty most frequently occurring combinations for each study area are given in Table 4. These top twenty structural combinations accounted for seventy-nine per cent of the houses studied in Parke County, but only fifty-three per cent of those in Vermilion County, The latter area is again seen to have a more diverse house stock with less conformity to simplistic structural modes. 14 TABLE 3, ROOF-TYPE ON MAIN UNIT Parke County Vermilio Number n County Number Per Cent Per Cent Roof-Type of Houses of Total of Houses of Total Front Gable 246 10 377 17 End Gables 1,197 51 765 34 Decorative Front-Centered Gable 65 3 47 2 Multiple Gables 444 19 506 22 Flat Hip 17 1 50 2 Peaked or Ridge Hip 308 13 381 17 Gambrel 8 10 Mansard 2 Other 72 3 127 6 2,357 100 2,265 100 SOURCE: Field Surveys. Narrow oblong houses with end gables were the most commonly occurring structural types accounting for thirty-one per cent of the houses in Parke County and twelve per cent in Vermilion County. The two-story variety is, of course, found in Kniffen's "I-house" (Fig. 5a) representa- 13 tive of the more prosperous farmer in the years prior to the Civil War. However, this structural type accounted for only three and four per cent of the Indiana and Illinois houses, respectively. Much more common were narrow oblong, end-gabled houses of one and one and one-half stories (Figs, 5b and 5c). Together they accounted for 15 TABLE 4. CO^[PARISON OF STRUCTURE TYPES IN PARKE AND VERMILION COUNTIES Per Cent Per Cent Parke County Vermilion County Structure Type Houses Houses Narrow Oblong, end gables, 1 story 19 5 Narrow Oblong, end gables, 1-1/2 story 9 3 Narrow Oblong, end gables, 2 stories 3 4 Narrow Oblong, front gable, 1 story 1 Narrow Oblong, front gable, 1-1/2 stories 1 Narrow Oblong, Dec. front-centered gable, 1-1/2 stories 1 Narrow Oblong, multiple gables, 1 story 3 1 Narrow Oblong, multiple gables, 1-1/2 stories 1 Narrow Oblong, peaked or ridge hip, 1 story 2 Wide Oblong, end gables, 1 story 6 6 Wide Oblong, end gables, 1-1/2 stories 5 5 Wide Oblong, end gables, 2 stories 1 Wide Oblong, end gables, 2-1/2 stories 1 Wide Oblong, front gable, 1 story 3 2 Wide Oblong, front gable, 1-1/2 stories 2 3 Wide Oblong, multiple gables, 1-1/2 stories 1 1 Wide Oblong, peaked or ridge hip, 1 story 2 2 Square, end gables, 1 story 2 Square, end gables, 1-1/2 stories 3 Square, peaked or ridge hip, 1 story 5 2 Square, peaked or ridge hip, 2 stories 1 2 Irregular Massing, end gables, 1 story 1 Irregular Massing, multiple gables, 1 story 5 3 Irregular Massing, multiple gables, 1-1/2 stories 5 3 Irregular Massing, multiple gables, 2 stories 3 Irregular Massing, multiple gables, 2-1/2 stories 1 Irregular Massing, front gable, 1 story 1 Irregular Massing, front gable, 1-1/2 stories . 1 Irregular Massing, peaked or ridge hip, 1 story 1 TOTAL 79 53 SOURCE: Field Surveys. 16 ^^(o) Figure 5: Structural Types: (a) I-house, (b, c, and d) I-cottages, (e and f) double-I cottages, (g) double-I house, (h, i, j, k) unnamed, (1) pyramidal cottage, (m) cube or cubic house, (n and o) square cottage, and (p and q) irregular massing. 17 twenty-eight and eight per cent of the houses in Parke and Vermilion Counties. These structural types, which I call "1-cottages," probably represent the mainstream of the American house building tradition, These forms Cone a slight variant of the other) were brought to North America by British, French, Dutch, and other settlers at the very beginning of the Colonial period. In the British experience, these structural types 14 derive from rural "cottages," The term cottage, previously applied by Glassie to a restricted type of New England house, seems appropriate here as a general descriptor. Narrow oblongs with more "exotic" roof styles were relatively rare in the two study areas . One exception was the one-story house with multiple gables not previously identified in the literature (Fig, 5d) . These houses accounted for three per cent and one per cent of the houses in Parke and Vermilion Counties, respectively. Generally, the more complex roof styles were associated with the wider and more variegated floor plans. No attempt was made to name these latter structures. Wide oblong houses with end gables were the second most common structural types, accounting for eleven and twelve per cent of the houses in Parke and Vermilion Counties. Houses of one and one and one-half stories, which I call "double-I cottages," predominated (Figs. 5e and 5f ) . To casual view, these latter structures may seem very similar in appearances to I-cottages, However, in both the two-story and two and one-half story varieties the scale differences derived from increased breadth or width in floor plan are more clearly evident (Fig. 5g) . These latter houses I have dubbed "double-I houses," This structural type is associated more with New England and the Mid-Atlantic states as well as 18 with portions of the Middle West settled originally by people from these latter areas. "New England Large" houses and the "Four over Four" houses of Pennsylvania identified by Pillsbury and Kardos, fall into this category as does Glassie's "Mid-Atlantic farm houses," The wide oblong with front gables accounted for five per cent of the houses of both study areas. Illustrated are the one and the one and one-half story varieties (Figs. 5h and 5i) . Multiple-gabled and peaked or ridge-hipped roofs in combination with wide oblong floor plans ac- counted for another three per cent of the houses in each study area. Only the single story varieties are shown (Figs. 5j and 5k). Until these structures are identified elsewhere and their origins clearly ascertained, it seems unwise to attempt to name them. However, it should be recognized that the various roof treatments distinguish these structures from double- I cottages. Distinguishing names should be developed. The square floor plan seems to have enjoyed its greatest popularity in the late nineteenth and early twentieth centuries. The square of one story was associated most frequently with a peaked or pyramidal roof (Fig. 51). The term "pyramidal cottage" seems appropriate for this particular structure. Five and two per cent of the houses in the Indiana and Illinois study area, respectively, were so characterized. The two- story variety is usually referred to as a "cubic" or "cube" house (Fig. 5m) . This house was often built from stock plans or even ordered out of catalogues. Square single or one and one-half story structures with end-gabled roofs characterized five per cent of the Parke County houses (Fig. 5n and 5o) , These I call "square cottages." In western Indiana they seem to be associated with marginally successful farms of the early 19 twentieth century or are located near small coal tovms of the same period. The term "square cottage" is suggested only because of the similarity to size to "I" and "double-I" cottages. Irregular massing, as previously described, was more of a miscel- laneous category and it clearly needs further refinement. The combinations of geometries produced by varied floor plans can provide a wide range of very different structural forms. However, roof -types do serve to partially offset or obscure these differences. This is particularly true of the multiple-gabled roof. The multiplicity of gables, designed to cover the integrated box-like spaces generated by an irregular floor plan, plays a major integrating role. Varying in height from one to two and one-half stories, these structures accounted for ten per cent of the houses in both study areas. Illustrative examples are shown (Figs. 5p and 5q) . In addition, four per cent of the houses in Vermilion County were irregularly massed with end gables, front gables, or peaked or ridge hip roofs attached. Whether names should be applied to any of these latter structural types is again debatable. I would prefer to see the particular forms of irregular massing identified before any nomenclature is applied. In the interim, a general designation such as "villa" might be applied to all irregularly massed houses. Most of these houses derive from the pattern books of the mid- to late-nineteenth century. Prepared by architects such as Andrew Jackson Downing, these architectural guidebooks 18 contained blueprints for rural houses or "villas." Most plans called for irregular massing in keeping with the ecletic tastes of the late- nineteenth and early-twentieth centuries. 20 ARCHITECTURAL STYLE CATEGORIES Seven categories were used to identify architectural decoration: federal, classical revival, gothic revival, Italianate, French American, and free classic. Their use on the houses of both study areas is given in Table 5. These style categories were based on Wilbur Peat's inter- 19 pretation of architectural decoration in the Middle West. Peat, an architectural historian, recognizes many differences between Eastern and Middle Western style applications, TABLE 5, ARCHITECTURAL STYLES Parke County Vermilion County Number Per Cent Number Per Cent Style of Houses of Total of Houses of Total Federal and Late Federal Classical Revival Gothic Revival Italianate Free Classic Other No Architectural Decoration 1,967 2,357 100 2,265 100 SOURCE: Field Surveys. Federal style Csometimes called "early republican" style) was most popular in the rural Middle West between 1810 and the Civil War. This style is characterized by brick construction, the geometric symmetry of 117 5 22 1 13 1 6 138 6 27 1 27 1 34 2 56 2 64 3 . 39 2 10 967 83 2,102 93 21 facade openings, low-pitched roofs, undecorated cornices under the eaves (which were often swept back at the corners and brought short distances across the end walls), windows left unframed (Fig. 6). This style gave an understated Image of refined order thought proper to the middle and upper classes of a newly formed republic. Relatively sober and plain. It was an artistic departure from the English Georgian modes thought 20 symbolic of colonial subordination. However, many of the houses so categorized In Parke County (five per cent) and In Vermilion County (one per cent) were of a later, but related, architectural mode characterized by severe simplicity In facade design and frame, as opposed to brick, construction. Accordingly, It was termed "late federal." Fig. 6: An I-House with Federal Architectural Styling. 22 The classical revival decoration, particularly Greek revival style, was popular in the Middle West from 1810 through to the 1950' s. Typical characteristics included low-pitched roofs which frequently ended in the pediment of a front gable, wide entablatures under the eaves, windows simply framed, and vertical corner boards (Fig. 7). Also characteristic was the use of classical columns and pilasters. The classical revival style was part of a strong nationalistic feeling; indeed, the nation came 21 to symbolize itself architecturally as classical democracy reborn. Applied first to public buildings in Philadelphia and Washington, D.C., it quickly diffused into the Middle West where it thrived especially in small towns and in the cities as domestic architecture. After the Civil War the style returned in a more elaborate and fanciful form often termed "classic resurgent" or "neo-classical revival." Approximately one per cent of the Parke County houses were found to have some kind of classical styling. The gothic revival decoration was characterized by steeply pitched roof lines, elaborate undereave decorations (especially barge boards with tracery), tall windows, and corner boards: all designed to accen- tuate vertical house lines as opposed to the horizontal emphases of federal styling (Fig. 8). Gothic revival architecture was popular between 1850 and 1900; however, this style, in understated form, was still used on houses constructed in rural Indiana and Illinois well after the turn of the twentieth century. It was the most prevalent architectural style in Parke County (six per cent of the houses) , but was less frequent in Vermilion County (one per cent) , Gothic revival styling in combination with irregular massing was widely popularized by landscape architects in pattera-books . 23 Fig, 7: A Two-Story Wide Oblong with Front Gable with Classical Revival Architectural Styling. Italianate styling, typified by low-pitched roofs often flattened over extended eaves, entablatures with conspicuous bracketing beneath the eaves, and tall windows capped with narrow round arches, often Included square campanile towers and cupolas (Fig. 9), Popular in the Middle West between the 1850' s and the 1880' s, Italianate was primarily, although not exclusively, an urban form and was particularly popular on commercial buildings. Known variously as "Anglo-Italian," "Italian villa," and "Tuscan villa," it was patterned after Italian medieval buildings and was characteristically informal and asymmetrical. It was frequently associated with irregular massing. One per cent of Parke 24 Fig, 8: An I-Cottage with Gothic Revival Decoration, The porch is a later addition. County *s rural houses and two per cent of those in Vermilion County were classified as Italianate. Another closely related style, the French- American mode (often called "Franco-American" or "French Second Empire") was not found in Parke County, Only three examples were found in Vermilion County, The styles commonly known as "neo-Jacobean," "Queen Anne Revival," and/or "free classic" were all treated under the single designation of 22 free classic (Fig, 10), Borrowing from and combining the attributes of a host of traditional architectural modes, free classic styles were almost always associated with the use of irregular massing. Decorative 25 Fig, 9: An Irregularly Massed House with a Flat Hipped Roof and Italianate Decoration. elements included variously shaped and sized gables, bays, dormers, towers, and porches (all delicately trimmed for an effect of structural lightness), variegated siding, prominent chimneys, and undereave decora- tion featuring simple cornices and bargeboards. Free classic style was found in two and three per cent of the houses in Parke and Vermilion Counties respectively. Houses with the older federal, classical revival, and gothic revival styles of decoration accounted for twelve per cent of Parke County's total houses. This contrasted sharply with the two per cent figure for Vermilion County. The latter area, however, did have a slightly larger 26 Fig. 10: An Example of "Free Classic" Decoration Where a Variety of Architectural Styles are Combined. In this instance the "Queen Anne" motif dominates. representation of houses with the more recent Italianate and free classic styling. Of greater significance was the general lack of architectural decoration in both study areas. Eighty-three per cent of the houses in Parke County had no architectural decoration whatever. This figure climbed to ninety-three per cent for Vermilion County. Clearly, verna- cular building ideas have dominated both areas. 27 HOUSE AGE In keeping with settlement geography's past emphasis on traditional or folk architecture, houses built before 1920 were isolated for further analysis. Rickert's dating system was employed to establish the age of 23 houses by general time period (Table 6). Rickert's system relies on an elaborate checklist of house characteristics which includes building materials, shapes and sizes of structural components, and architectural features among other criteria. Fifty-eight per cent of the houses in Parke County were estimated to have been built before 1920. For Vermilion County this figure fell to thirty-seven per cent. Only three per cent of the houses in the Indiana study area and ten per cent in Illinois could not be dated. TABLE 6. ESTIMATED AGE OF HOUSES BY PERIOD ' Parke County Vermilio Number n County Number Per Cent Per Cent Period of Houses of Total of Houses of Total Prior to 1850 59 3 8 1850 to 1920 1,306 55 844 37 1920 to 1945 519 22 652 29 1945 to 1970 409 17 535 24 Date undetermined 64 3 226 10 2,357 100 2,265 100 SOURCE: Field Surveys, 28 HISTORICAL HOUSE TYPES Older houses were divided into two categories (1) pre-1850 houses and (2) houses built between 1850 and 1920. Their structural and architectural characteristics were combined to identify historical house types. Specifically, I was interested to see how the various structural types associated with the various architectural styles for each historical time period. I decided to focus on only the most frequently occurring historical house types and sought to limit attention to those house types which accounted for at least one per cent of the total houses in each respective study area. Unfortunately, no house type in either area accounted for even this small a proportion. The cutoff level was then reset at one-half of one per cent. Only seven historical house types were found in Parke County (Table 7) and none in Vermilion County. In the aggregate, the historical house types identified comprised only five per cent of the Parke County houses studied. The "I-house" (Fig. 11a) and "I-cottage" (Fig. lib) with federal or late federal styling characterized the pre-1850 period. As distinc- tive survivors of the early period of settlement they are the most numerous historical house types in the Parke County landscape, Gothic style dominated the houses of the post-1850 period. The one and one- half story I-cottage with end gables (Fig. lie) and the I-cottage with front-centered gable (Fig. lid) appeared. Also appearing was a one and one-half story narrow oblong with front gable (Fig, lie). The term "shotgun house" has previously been applied by Lewis to houses of this 24 structural type. One and one-half story narrow oblongs, one with a 29 peaked or ridge-hipped roof (Fig, llf) and the other with multiple gables (Fig. llg) , also appeared. I have coined the terms gothic revival modified I-cottage and gothic revival multi-gabled I-cottage to apply to these two latter house types. TABLE 7. HISTORICAL HOUSE TYPES SURVIVING IN THE PARKE COUNTY LANDSCAPE^ Pre-1850 1850-1920 Federal I-house * Federal I-cottage * Gothic Revival I-cottage (end gables) * Gothic Revival I-cottage (end and front-centered gables) * Gothic Revival Shotgun * Gothic Revival Pyramidal I-cottage * Gothic Revival Multi-gabled I-cottage * SOURCE: Field Study. Includes only those house-types which accounted for at least one-half of one per cent of the total houses studied. It is disappointing that the house types identified as historically significant were not more prominent numerically. Again, only five per cent of the total houses could be grouped into seven historical house type categories. This suggests the extent to which houses of distinc- tive house type characteristics have disappeared from the rural landscape and the extent to which more recent house construction has come to dominate the rural scene. The settlement geographer may well 30 continue to focus on traditional folk house types, but he should fully realize that he may be dealing with an exceptionally small proportion of the total house stock available for study. Fig, 11: Historical House Types Significant in Parke County, Indiana: (a) Federal I-house, (b) Federal I-cottage, (c and d) Gothic revival I-cottages, (e) Gothic revival shotgun house, (f) Gothic revival modified I-cottage, and (g) Gothic revival multi-gabled I- cottage. Of course, the student of historically significant house types should also realize that houses have always been highly personalized 31 buildings. Whereas certain structural and style trends can be discerned, their components have always been combined to produce highly ecletic buildings. Ecleticism in houses became increasingly important through the late nineteenth century. For example, there was a definite trend away from simplistically structured and styled I-houses and I-cottages toward houses which used varied floor shapes and roof types along with stylistically complex architectural decorations. CONCLUSIONS Although geographers have long been interested in identifying house types and in tracing their origins and diffusion, they know very little about how houses with different characteristics vary in the contemporary scene from place to place. Settlement geographers interested in houses have but recently arrived at the classification stage in their work. They have yet to agree as to which house characteristics are really significant and how these characteristics should be scaled. They have yet to develop adequate classification techniques. This study has proceeded from a set of simplifying assumptions about houses and how people view and remember them. Each of these assumptions deserves detailed analysis. Can structure and decoration be separated? If so, is architectural style really as significant to house appearance as structural form? Do main house units deserve special focus or should extensions be automatically included in analysis? We need agreement as to what constitutes the principle elements of house structure. Do people really tend to view house structures in terms of floor shape, building height, and roof line? Whatever cognitive 32 dimensions Americans use in evaluating houses, these ought to be built into comprehensive categorization schemes. Although future researchers may reject the assumptions upon which this study has been based and will certainly apply more sophisticated sorting techniques, the house typing scheme used here has proven adequate to establishing a number of important facts. First, it has identified a most significant American house form; the use of one and one and one- half story narrow and wide oblong floor plans with end-gabled roofs. These houses I have named "I" and "double-I" cottages and have systemati- cally determined their proportion in two sample areas. Certainly this is only a start toward determining what typifies the American landscape so far as house structure is concerned. I would hypothesize, however, that "I" and "double-I" cottages will prove to be dominant house forms in most portions of the United States. Second, when applied to pre-1920 houses, the classification procedure used was able to identify a limited number of historical house types. This endeavor serves to remind us how little remains of the geographic past in the contemporary landscape. Analysis suggests that students in focusing on folk, vernacular, or common house types of the nineteenth and early twentieth centuries may be restricting themselves in the Middle West to as little as five per cent of the total houses even where conditions have favored the survival of old houses. The third accomplishment relates to what we have called architec- tural style. Architecture, I maintain, combines notions of both structure and style. Emphasis in the past has been on style or decora- tion. If the two counties inventoried are typical of the Middle West, 33 then only about ten per cent of the region's rural houses might be expected to have clear signs of architectural decoration. This suggests that emphasis should be placed on identifying basic structural types as has been the trend in settlement geography. Architectural historians should be encouraged to divorce structure as an aspect of style and treat it as a phenomena with its own independent architectural signifi- cance. Once basic structural types are understood, then study of their association with stylistic modes should be pursued vigorously. Finally, in offering a comparison of rural houses in two Middle Western counties, a house typing system was used which first treated individual house characteristics; characteristics were then combined to identify structure types and, with the inclusion of architectural style, house types. This system counters the traditional method of evaluation where houses with distinctive characteristics are classified in the field according to previously determined house type categories. Floor shape, height, roof type, architectural style, and estimated age characteristics were shown to vary in ways consistent with the historical and contemporary images of the two areas. A much elaborated system should be developed and applied to the comparative analysis of house types across the United States. Some systematic form of house typing is needed if we are to move beyond the stage of subjective eyeballing. If we are to objectively compare the houses of different areas, analyze their spatial distributional patterns in search of origins, and understand their survival in the present day, then an objective and universally appointed means of identifying house types seems in order. Once settlement geographers 34 have agreed upon the criteria by which house architecture is to be evaluated, then attention should be turned to applying sophisticated sorting or categorization techniques. Preferably, such classification would expand upon existent house type nomenclature in order to build firmly on the existing literature. FOOTNOTES Fred Kniffen, "Folk Housing; Key to Diffusion," Annals, Association of American Geographers, Vol. 55 (1965), pp. 549-577 and "Louisiana House Types," Annals , Association of American Geographers, Vol. 26 (1936), pp. 179-93; Henry Glassie, Pattern in the Material Folk Culture of the Eastern United States (Philadelphia: University of Pennsylvania Monographs in Folklore and Folklife No. 1, 1969, 1968); Peirce Lewis, "Common Houses, Cultural Spoor," Landscape , Vol. 19 (1975), pp. 1-22 and "The Geography of Old Houses," Earth and Mineral Sciences , Vol. 39 (1970), pp. 33-37; Richard Pillsbury and Andrew Kardos, A Field Guide to the Architecture of the Northeastern United States (Hanover, NH: Dartmouth University, Geography Publications at Dartmouth, No. 8, n.d,). 2 Kenneth E. Corey, "A Spatial Analysis of Urban Houses," Unpublished Ph.D. dissertation. University of Cincinnati, 1969. 3 A bibliography of the house type literature in geography appears in John A. Jakle, Past Landscapes; A Bibliography for Historic Preservationists Selected from the Literature of Historical Geography (Monticello, IL: Council of Planning Librarians, Exchange Bibliography 651, 1974). 4 John E. Rickert, "House Facades of the Northeastern United States; A Tool of Geographic Analysis," Annals , Association of American Geographers, Vol. 57 (1967), pp. 211-238. Research on how Americans stereotype house form has just begun. See: Bernard Pyron, "Form and Space Diversity in Human Habitats, Perceptual Responses," Environment and Behavior , Vol. 3 (1971), pp. 382- 411 and Donald Appleyard, "Why Buildings are Known, A Predictive Tool for Architects and Planners," Environment and Behavior , Vol. 1 (1969), pp. 131-156. The literature on traditional decorative styles in the United States is extensive. The best summaries include: Alan Gowans, Images of American Living: Four Centuries of Architecture and Furniture as Cultural Expression (Philadelphia: J.B. Lippincott, 1964) and Marcus Whiff en, American Architecture Since 1780, A Guide to the Styles (Cambridge, MA: The MIT Press, 1969). The literature on style percep- tion is less complete. For the best statement see: Russell Lynes, The Tastemakers (New York: Harper and Brothers, 1949). Two recent articles by geographers show special concern for decorative styling. See: Lewis, "Common Houses, Cultural Spoor," op. cit. , footnote 1 and Robert Bastian, "Architecture and Class Segregation in Late Nineteenth-Century Terre Haute, Indiana," The Geographical Review , Vol. 65 (1975), pp. 166-179. Q This generalization has been substantiated in an investigation of how children see and remember houses, George Rand, "Children's Images of Houses; A Prolegomena to the Study of Why People Still Want Pitched Roofs," in William J. Mitchell (ed.). Environmental Design: Research and Practice (Los Angeles: University of California, 1972). I 36 9 Lewis, "Coiranon Houses," op. cit ., foot note 1, p. 15. Wilbur D. Peat, Indiana Houses of the Nineteenth Century (Indianapolis: Indiana Historical Society, 1962), p. 121, Kniffen, "Folk Housing," op. cit ., footnote 1, p, 560, 12 Pillsbury and Kardos, op. cit ., footnote 1, p. 27 and p. 55. 13 Kniffen, "Folk Housing," op. cit ., footnote 1, p. 550, 14 The evolution of the cottage is discussed in: lorwerth C. Peate, The Welsh House, A Study in Folk Culture (London: Society of Cymmrodorian, 1940). Also see: V»M« ^^id F.J. Chesher, The Cornishman's House (Truro: D. Bradford Barton Ltd., 1968) and Sidney 0. Addy, The Evolution of the English House (London: Swan Sonnenschein and Co., 1898). Glassie, op. cit ., footnote 1, p, 131. 16 Pillsbury, op. cit ., footnote 1, p. 27 and p. 55 and Glassie, op. cit . , footnote 1, p. 58. Rickert, op. cit ., footnote 4, p. 229. 18 Perhaps, the most widely read architectural guide was Andrew Jackson Downing, Cottage Residences (New York: William and Putnam, 1842). 19 Peat, op. cit ., footnote 10. ^ ^Ibid ., p. 9. 21 See Talbot F. Hamlin, Greek Revival Architecture in America (New York: Oxford University Press, 1944). 22 Peat, op. cit ., footnote 10, p. 149. 23 Rickert, op. cit ., footnote 4, p. 216. 24 Lewis, op. cit ., footnote 4, p. 36. Occasional Publications Department of Geography University of Illinois Publ ished Paper s A Theoretical Framework for Discussion of CI imatological Geomorphology, by Dag Nummedal, April, 1972. Social Areas and Spatial Change in the Black Community of Chicago: 1950-1960, by Charles M. Christian. April, 1972. Regional Components for the Recognition of Historic Places, by Richard W. Travis. October, 1972. 1, k Matrix and Graphic Solutions to the Traveling Salesman Problem, by Ross Mullner. October, 1972. 5 Regional Changes in Petroleum Supply, Demand and Flow in the United States: I966-I98O, by Ronald J. Swager. April, 1973. I. 6 Social Problems in a Small Jamaican Town, by Curtis C. Roseman, Henry W. Bullamore, Jill M. Price, Ronald W. Snow, Gordon L. Bower. April, 1973. Some Observations on the Late Pleistocene and Holocene History of the Lower Ohio Valley, by Charles S. Alexander. April, 197^. Methods and Measures of Centrography : A Critical Survey of Geographic Applications, by Siim Soot. April, 1975. A Re-Evaluation of the Extraterrestrial Origin of the Carolina Bays, by J. Ronald Eyton and Judith I. Parkhurst. April, 1975. Glacial Maximum Tundra: A Bioclimatic Anomaly, by Joseph M. Moran. August, 1976. 11 A Comparative Analysis of Rural Houses in Twc Middle Western Counties; The Testing of a House Typing System, by John A. Jakle. August, >976.