Importance of the Task: Low or High Stakes
One aspect of external pressure concerns whether the task is perceived as important, specifically whether it is viewed as a low- or high-stakes requirement. In a low-stakes, relaxed task, there is less stress during the task. In a high-stakes task or set of tasks, such as those found on an English competency examination for graduation or for university entrance, much more anxiety can be expected. Those learners who tend to be anxious anyway may become particularly tense while doing a high-stakes task. Skehan (1996a) discussed the differential effects of low- and high-stakes tasks.

Timing
The amount of time allotted for the task can be a major factor (Honeyfield, 1993; Skehan, 1996a), especially for L2 learners who are at the beginning and low intermediate levels. When a task is "speeded," that is, when only a certain amount of time is given to complete the task, it might become more difficult for some learners. If students are allowed to take all the time they need, i.e., if the task is "unspeeded," this takes off some of the pressure. In-class tasks do generally have a time limit, although, depending on the task type and the goals, some tasks that are unfinished can be done as homework assignments.

Input Genre and Modality
Tasks can be analyzed according to the input genre (newspaper article, diary, recipe, diary, TV show, conversational talk, lecture, and so on) and modality (e.g., written, spoken, graphic/pictorial) (Honeyfield, 1993; Nunan, 1989; Skehan, 1996a). Genre and modality interact. For instance, a newspaper article can be a written text and an accompanying picture, and it can also be read aloud.

Richards and Rodgers (2001) cited a range of input materials for L2 tasks, including books, newspaper, video, TV, and so on. Interest level of the learners in the material is particularly crucial. If materials are perceived as boring or as too easy or too difficult, learners will be unmotivated to do the task (Scarcella & Oxford, 1992). Publishers provide materials of wide interest to most students, although cultural factors such as religion can prevent some materials from being used for L2 tasks in particular locations.

Also, relevance and suitability of task input-and of tasks themselves-also depend on whether the L2 learning occurs in a foreign versus a second language setting. Certain input and tasks would be more available and feasible in a second language environment than a foreign language environment, because in the former there are many more natural resources in the target language and many more native speakers of the language with whom to interact. Yet because of the Internet, the foreign language environment now contains instant L2 input (not just written text, but also multimedia that could help develop multiple skills) that were simply unavailable to learners in times past. In locations where students have easy access to the Internet, teachers can take advantage of new input in simulations and WebQuests. The widespread presence of games and videogames on the Internet creates additional input possibilities. However, in some Asian countries, many learners are already so involved in L1 videogames for entertainment that they might not recognize L2 game-based or videogame-based tasks as a serious endeavor. The context determines the relevance of various types of input.

Linguistic Complexity
An important task factor is linguistic complexity (Dahl, 2004; Foster & Skehan, 1996; Richards & Rodgers, 2001; Skehan, 1996a), such as number of words in a sentence, amount of redundancy, degree of use of dependent clauses and other complexity-creating structures, discourse style, sequence complexity, technicality of vocabulary, concreteness or abstractness, sectioning, and other features. As noted by Dahl (2004), linguistic complexity is not synonymous with "difficulty" but is instead an objective property of a system-a measure of the amount of information needed to describe or reconstruct it. It is the result of historical processes of grammaticalization and involves mature linguistic phenomena (Dahl, 2004). Gibson (1998) indicated that linguistic complexity is a function of the "integration cost" and the "memory cost" associated with keeping track of obligatory syntactic requirements, such as center-embedded dependent structures, placement of large phrases earlier (heaviness effect), and ambiguity effects.

Salaberry (2001) mentioned the following issues involved with task language features: frequency and saliency; and linguistic categories, such as vocabulary, phonology and phonetics, morphosyntax, discourse, pragmatics/speech acts, and sociolinguistics. All of these contribute in various ways to the degree of linguistic complexity.

Linguistic complexity is not the same as "difficulty." The person's familiarity with the material, the topic, or the language properties mitigates some of the difficulty even when the linguistic material is complex. The difficulty is also affected by the number of language skills (reading, writing, speaking, and listening) and subskills required to do the task
Cognitive Load and Cognitive Complexity
Cognitive load is another feature of the task. The concept of cognitive load relates to Sweller's (1988, 1999) assumption that people's capacity to process information is limited. The more that a learner tries to hold in his or her head at a given moment, the harder the learning is and the more likely there will be a cognitive overload. Another assumption is that some tasks have a higher cognitive load. For instance, the task of integrating information from multiple sources might have a higher cognitive load than the task of following an example. Cognitive load can be increased by competing stimuli in the input or during the task, distracting the learner.

Cognitive complexity is yet another characteristic, but it relates not just to the task but also to the person. Analysis of cognitive complexity has been defined as "an aspect of a person's cognitive functioning which at one end is defined by the use of many constructs with many relationships to one another (complexity) and at the other end by the use of few constructs with limited relationships to one another (simplicity)" (Pervin, 1984, p. 507). Therefore, cognitive complexity involves a person component (unobservable cognition and observable behavior) and a task structure component. If a computer is involved, there is also an interactive system component (Rauterberg, 1992).

The task-required cognitive processing operations can be complex (Foster & Skehan, 1996; Ikeda & Takeuchi, 2000; Richards & Rodgers, 2001; Skehan, 1996a), but not every cognitively complex task is viewed as difficult. Whether or not a particular student actually perceives a given, cognitively complex task to be difficult and challenging depends considerably the student's familiarity with the kind of cognitive operations required.

Interaction and Output Demands
Presence or absence of a demand for output is a task factor. Swain (1985) and Scarcella and Oxford (1992) emphasized the importance of students' providing comprehensible output in task situations, often through interaction with others. Task interaction may be one-way, as in one person talking and the other listening or writing notes. It may be two-way (Long, 1985; Richards & Rodgers, 2001), as in two individuals engaged in an information-gap task (Doughty & Pica, 1986; Nunan, 1989) or sharing personal experiences (Foster & Skehan, 1996). It may be multi-way, as in a group discussion, role-play, or simulation (Crookall & Oxford, 1990). Among many examinations of which types of tasks promote L2 learning (see, e.g., Plough & Gass, 1993; Robinson, 1995; Yule et al. 1992), a review by Pica et al. (1993) reported that negotiation of meaning is most likely to occur when learners are involved in an interaction with the following four features:

*Each of the students holds a different portion of information that must be exchanged and manipulated in order to reach the task outcome.
*Both students are required to request and supply this information to each other.
*Students have the same goal.
*Only one outcome is possible from their attempts to meet the goal.
Thus, qualitative differences in the nature of the negotiation of meaning resulting from different tasks and different types of interaction, as Nunan (2004) also pointed out.

However, interaction and output might not be essential, depending on the task purpose. For learning the use of relative clauses, Tanaka (1996, in Ellis, 2003a) found that practicing with input proved to be more efficient than practicing with output (using relative clauses in traditional production-practice tasks). Input practice tasks helped students understand relative clauses better, and their ultimate production ability was just as strong with input practice tasks as with traditional production-practice tasks.

When production practice is the goal of the task, complexity of the output becomes a task factor. Output complexity relates to the complexity of language the learner uses and the cognitive sophistication of the output, both of which depend on the learner's willingness to take risks in restructuring forms and concepts (Foster & Skehan, 1996; Skehan, 1998b).

Allowable Amount of Planning
The amount of planning (a metacognitive learning strategy; see Oxford, 1990) allowed or encouraged is a factor in how well the learner accomplishes the task. Foster and Skehan (1996) examined the influence of task type and degree of planning on three different aspects of L2 performance: fluency, accuracy, and complexity. The study employed three types of tasks (personal information exchange, narrative, and decision-making) under three planning conditions (unplanned, planned but without detail, and planned with detail). Results indicated that planning had clear effects on both fluency and complexity of participants' output. However, planning was not the key to accuracy. In fact, less detailed planners were more accurate than non-planners and those who planned in detail. Interactions emerged between task type and planning conditions. Effects of planning were greater with narrative and decision-making tasks than with personal information exchange tasks. In their discussion, Foster and Skehan noted that a trade-off existed between the goals of performance complexity and performance accuracy. They explained that individuals have a limited capacity for attention, as noted earlier, so when a task is more cognitively demanding, attention is diverted from formal linguistic features-the basis of accuracy-to dealing with these cognitive requirements.

Sometimes when learners are allowed an opportunity to plan, this makes the task seem easier, but at other times the allowance of planning sends a signal that this is a difficult task, which makes certain learners anxious. The way the planning is introduced and implemented influences the value of planning.

Timing
The amount of time allotted for the task can be a major factor (Honeyfield 1993; Skehan 1996a), especially for L2 learners who are at the beginning and low intermediate levels. When a task is "speeded," that is, when only a certain amount of time is given to complete the task, it might become more difficult for some learners. If students are allowed to take all the time they need, i.e., if the task is "unspeeded," this takes off some of the pressure. In-class tasks do generally have a time limit, although, depending on the task type and the goals, some tasks that are unfinished can be done as homework assignments.

Teacher Roles and Characteristics
Teachers can take many different roles in regard to L2 tasks (Honeyfield, 1993; Nunan, 1989; Oxford, 1990; Scarcella & Oxford, 1992; Willis, 1996a, 1996b, 1998). Richards and Rodgers (2001) and Scarcella and Oxford (1992) mentioned the following task roles for teachers: selector/sequencer of tasks, preparer of learners for task, pre-task consciousness raiser about form, guide, nurturer, strategy-instructor, and provider of assistance. Cultural and linguistic backgrounds and teaching styles influence the roles teachers feel comfortable taking (Oxford, 2002; Oxford, Massey, & Anand, 2003; Scarcella & Oxford, 1992). The amount and kind of help provided by the teacher was singled out as a task-related teacher factor by Honeyfield (1993) and Scarcella and Oxford (1992).