6.1 Defining the study areaIn order to pursue an empirical study in geography one has to define one's study area. However, unlike regions, places are not defined by their boundaries but by their focus. This gives them a kind of point-like character. But instead of being mere points, the influence of places extends from the central focus towards the landscape around it. Defining definite borders of place is therefore usually impossible and seldom necessary. For the methodological purposes, however, it is necessary to define some kind of study area where to collect the empirical data.
In the case of Roseau the areal extension of the city is not self-evident. Often, only the Inner City is called 'Roseau'. The district of Roseau City Council, on the other hand, extends several kilometres further both northwards and southwards (see Chapter 7 for further discussion). My primary study area is the continuous urban area of the city, mostly framed by steep hillsides (Fig. 5).
The study area is approximately the area covered by sheets 6490 and 6491 of Dominica 1:2500 map series (Dominica 1:2500) except for the easternmost parts of sheet 6490 that cover Upper Kings Hill and Morne Bruce areas.
In the north, the area reaches up to the northern end of Goodwill so that that the harbour and warehouse district of Woodbridge Bay is excluded, as well as Stock Farm which is situated on the top of a hill. In the south the study area reaches till Newtown, and lower parts of Lower Kings Hill are also included. In the west the Caribbean Sea forms a natural border while in the east the border is more obscure. East from Upper Goodwill there is the residential area of St. Aromant, and urbanity lessens gradually eastwards. My study area ends some hundred meters east from the crossroads of Federation Drive and the way between Bath Estate and St. Aromant. East from that point the buildings are distributed in a relatively sparse manner.
I call the sub-divisions of the study area 'districts'. To avoid conceptual confusions, I do not use the term 'place' for them. As they are relatively small areas, however, the concept of 'region' is not appropriate. I also want to avoid the arid abstractness of the concept of 'area'. Term 'district' is also used by Lynch (1960) in his theory about urban imageability.
Inside the convex hull of the study area there are also the districts of Bath Estate, Botanical Gardens, Morne Bruce, and Kings Hill. Of these, I have included Bath Estate and Botanical Gardens in the study area. Bath Estate, while not part of the continuous urban area in a strict sense, is still urban in character, and separated from the rest of the city by Botanical Gardens and Roseau River. Botanical Gardens are situated just next to Cental Roseau, and are usually seen as part of the city (see Chapter 12.3). Morne Bruce and Kings Hill, on the other hand, are sparsely built districts up on the hills. For practical reasons, they were excluded from the study.
I had roughly divided the study area into six districts before the field work period in Roseau. The division was based on cartographic data and on the experiences of my previous visit to the city in autumn 1993. The districts were Central Roseau, Potter's Ville, Goodwill, Bath Estate, Botanical Gardens, and Newtown. My method of testing this hypothesis was that I asked the people who were in the streets around my hypothetical border area to which district the place where we were belonged to, and asked them to tell where the border line was. The border between Central Roseau and Potter's Ville was taken to follow the course of Roseau River, and the border between Central Roseau and Botanical Gardens was taken to be the wall of the gardens, and on them I posed no questions.
Two changes were made to my original division. The border between Goodwill and Potter's Ville was moved down to the west were the slope begins to rise up to Goodwill. Originally, I had marked the border on the east side of Church Lane where an unbuilt area is situated. This presupposition was proved wrong by the people in the street. Another change was that an area that I had presupposed to be part of Newtown was separated as its own district of Lower Kings Hill.
Most ambiguous of the border lines between the districts is the one between Central Roseau and Newtown. People largely disagreed if the border was on the top of the hill where State House is situated--my original proposal--or further south-east on the edge of Newtown Savannah. During the study period I asked several inhabitants who I met on the streets between State House and the Newtown Savannah where the border between Newtown and Roseau Downtown was. Out of twelve of those who could answer, seven favoured the State House hill while five proposed some other delinement closer to the Savannah. Thus, I chose State House hill to be the border.
Most of the districts were further divided into two smaller sub-districts (see Fig. 5). The borders of them are only approximate and rather methodological in character.
Terms 'Inner City', 'Central Roseau' and 'Downtown' are used in this study so that 'Inner City' refers to the traditional area of Roseau, i.e. it covers the districts of Central Roseau and Botanical Gardens (with Windsor Park). 'Central Roseau', on the other hand, is the Inner City grid area reaching from Roseau River in the NW to the hill that separates Newtown from the Inner City in the SE, excluding the district of Botanical Gardens. 'Downtown' is a sub-division of Central Roseau and refers to the central business district of the city (excluding the French Quarter; see Fig. 5).
Conceptual division between 'Central Roseau', 'Inner City' and 'Downtown' was done after the field study period. In the interviews term 'Roseau Downtown' was used to signify the whole district of Central Roseau. The districts under consideration were pointed to the interviewees on a Roseau map in scale 1:10,000 (Roseau insert in Dominica 1:50,000) and, when necessary, explained with words.