organizations that control critical resources. Thus, the greater the resources of an actor, the greater the power possessed by the actor, because a variety of resources can be used to negotiate the exchange. Therefore, power is vested in the number and availability of valued resources that may be used as concessions to influence another actor.
Another attribute related to power in the exchange relation is cohesion (Emerson 1962, 1972). Cohesion is defined as the average of two actors' power or dependence on each other. It represents a variable state of the exchange relationship (for example, unchanged, undergoing change, or extinguishing) and provides a measure of the nature of interdependence between the actors. Cohesion "comes into play whenever one or both actors in the relation encounter value conflict. In the absence of conflict at a given time, it is an assessment of the level of potential conflict which the relation can survive or absorb" (Emerson 1972:63). Thus, by definition, power and cohesion are directly related. Therefore, it is implicitly assumed that greater cohesion is related to greater satisfaction with the consequences of exchange (that is, perceptions of tourism impacts).
With power as a dimension of exchange between two actors, four outcomes (depending upon the level of each actor's power) are possible. A comparison of an individual actor's outcome from a disadvantaged or advantaged position provides the basis for determining an actor's perception of the exchange. This is shown by the matrix presented in Figure 2. Balanced exchange representing high and low levels of power held by the actors are found in quadrants 1 and 4, respectively. These two quadrants provide a basis for comparing actors' outcomes as either positive or negative. When power levels are high for both actors, exchange is mutually beneficial when compared to actors with low power levels, because the former find exchange rewarding. When the actors have an array of resources to exchange, there is a greater likelihood for exchange occurring as the actors can obtain at least one of a number of valued resources from the relationship and derive some benefit from it.
It is also possible for both actors to possess low levels of power (quadrant 4). This occurs when following an exchange, the consequences do not produce its expected benefits. The resources that have been made available and are used in the exchange have little or limited value. This outcome is described as unrewarding, since both actors derive little benefit from it. The prediction of an outcome for a balanced exchange relation is dependent upon comparison between the two possible states (that is, in terms of high or low levels of power).
Quadrants 2 and 3 of the matrix represent unbalanced exchange where power levels between the actors in the dyad are different. Depending upon the power position of the actor in relation to the other, exchange may be either advantageous or disadvantageous. It is advantageous for the actor with greater power, as the actor is able to obtain an additional valued resource(s) at the expense of the other actor who did not obtain the valued resource that was expected. While for the disadvantaged actor, who has a lower level of power, the exchange is seen as unfair and the actor is placed at a disadvantage. The power-dependence relation framework proposed by Emerson