A Computational Theory of Complex Problem Solving Using the Vector Space Model (part I): Latent Semantic Analysis, Through the Path of Thousands of Ants.
José Quesada, Walter Kintsch
Institute of Cognitive Science
University of Colorado, Boulder
Muenzinger psychology building
Campus Box 344
Boulder, CO 80309-0344
{quesadaj, wkintsch}@https://www.doczj.com/doc/4517646641.html,
Emilio Gomez
Department of Experimental Psychology
University of Granada
Granada (Spain)
egomez@ugr.es
Abstract. For years, researchers have argued that Complex Problem Solving
(CPS) is plagued with methodological problems. The interest of this research
paradigm, a hybrid between field studies and experimental ones, is tied to the
success of methodological advances that enable performance to be analyzed.
This paper introduces a new, abstract conceptualization of microworlds
research based on two theoretical lines: (1) a representational problem, where
protocols can be seen as objects in a feature space and, (2) a similarity measure
problem, where a similarity metric has to be proposed. To materialize this
conceptualization we introduce Latent Semantic Analysis (LSA), a machine-
learning model that induces representations of the meaning of words by
analyzing the relation between words and passages in large bodies of
representative text, and describe how LSA can be implemented as a theory and
technique to analyze performance in CPS, using actions or states as units
instead of words and trials instead of text passages. Basic examples of
application are provided, and advantages and disadvantages are discussed.
Many real-word decision making and problem solving situations are (1) dynamic, because early actions determine the environment in which subsequent decision must be made, and features of the task environment may change independently of the solver’s actions; (2) time-dependent, because decisions must be made at the correct moment in relation to environmental demands; and (3) complex, in the sense that most variables are not related to each other in one-to-one manner. In these situations, the problem requires not one decision, but a long series, and these decisions are, in return, completely dependent on one another. For a task that is changing continuously, the same action can be definitive at moment t1 and useless at moment t2. However,
traditional, experimental problem solving research has focused largely on tasks such as anagrams, concept identification, puzzles, etc. that are not representative of the features described above.
In Europe, a movement of researchers led by Broadbent (e.g., [2]) in the UK and D?rner (e.g., [8]) in Germany were concerned about that fact and started working on a set of computer-based, experimental tasks that are dynamic, time-dependent, and complex, called Microworlds1. This branch of the thinking and reasoning psychology has been called Complex Problem Solving (e.g., [10]).
Compared to traditional Problem Solving, Complex Problem Solving (CPS) radically changed the kind of phenomena reported, the kind of explanations looked for, and even the kind of data that were generated. However, the results obtained to date are far from being integrated and consolidated. This fact led Funke to affirm that ‘Despite 10 years of research in the area, there is neither a clearly formulated specific theory nor is there an agreement on how to proceed with respect to the research philosophy. Even worse, no stable phenomena have been observed’ ([11], p. 25). Almost another 10 years after Funke’s argument, although more empirical research has been conducted in the area, we cannot say that the situation has changed drastically. At this moment, there is no theory able to explain even part of the specific effects that have been described or how they can be generalized.
A theory of generalization and similarity is as necessary to psychology as Newton's laws are to physics [30]. However, for CPS there is no common, explicit theory to explain why a complex, dynamic situation is similar to any other situation or how two slices of performance taken from a problem solving task can possibly be compared quantitatively. Intractability issues have been raised [23]. This lack of formalized, analytical models is slowing down the development of theory in the field. At least two problems make it difficult to apply the classical problem solving approach to CPS:
(1) The utility of state space representation for tasks with inner dynamics is reduced because in most CPS environments it is not possible to undo the actions. For example, imagine that two participants in Firechief [27] are in an identical situation (system state) when the trial starts. One of them proceeds to make a control fire on the right side of a central fire, while the other one is preparing a control fire on the north front of the fire. After 20 seconds have elapsed, the system state is no longer identical for them. Now they have to cope with rather different problems. Moreover, if the first participant wants to apply the same technique that the second participant used, there is absolutely no way to come back to the initial state and begin with a new strategy. This situation is not an issue in static tasks like the tower of Hanoi problem because the last state is always available to ‘undo’ the wrong actions. Feedback delays (e.g., [1], 1 This term sometimes has other meanings. For example, educational applications created to teach physics [16], simulated words in the early AI programs like the block word of SHRDLU, [34] and static tasks to study decision making [14] have been called Microworlds. However, we are interested only in the tasks that fulfill the conditions described above.
[12]) and upsettingly large number of possible states (e.g., [8], [27]) contribute to the reduced utility of the state space approach.
(2) Traditional techniques of knowledge elicitation seem not to be very suitable: Concurrent verbal protocols consistently interfere with performance [7]; measures based on relatedness judgments like rating correlations or pathfinder distance correlations are not sensitive to context manipulations in naturalistic task like fire fighting [4].
In this paper we introduce a theory and methodology for Complex Problem Solving tasks based on Latent Semantic Analysis (LSA, [24]). The theory addresses issues concerning induction, representation, and application of knowledge. The main idea of LSA is that most knowledge areas contain enormous numbers of weak interrelations that can be used to infer more knowledge than that resulting from simple addition.
In the new LSA perspective, we are not depicting all the possibilities of the system (system’s state space), but only the paths that people have actually followed when interacting with it. This offers a realistic view of how the system is understood and used by humans. LSA is a computational theory on how environmental constraints are learned and how they can be described. Using the classical Simon’s parable of the ant and the beach ([31], p. 63), LSA would be describing and inferring the shape of the beach using the constraints that thousands of ants’ paths impose n each other. In this sense, LSA can be conceived as a computational extension to theories for describing environmental constraints, such as the abstraction hierarchy (AH, [29], [33]; see also Quesada, Kintsch and Gomez, this issue).
LSA has several interesting features that make it a suitable technique to analyze performance on a complex, dynamic task:
(1) It does not assume independence of decisions; indeed, it uses dependencies between decisions to infer structure. Some methods employed in the past treated CPS performance in a way that assumed that decisions are independent or have ‘short term dependencies’ only. For example, when transitions between contiguous actions are used as the unit of analysis, the method assumes that the only dependency of interest is the one between an action (or state) and the following one (e.g. [17], [27]).
(2) LSA reduces the dimensionality of the space. Imagine a hypothetical Problem Solving task that, when performed from the beginning to the end in one of the N possible ways, traverses 300 states. To make it a really simple example, let us assume that every state is described using 6 dichotomous variables (2^6 = 64 possible states). Since we have 300 states in our sample of performance, there are 64 ^ 300 = 7e541 possible paths in this task. Every sample would be represented as a matrix of 6 x 300 = 1800 values. With LSA, every sample is represented as a vector of only 100- 300 values.
(3) There are no a-priori assumptions about ‘the beach’. To continue our previous example, not all of the 7e541 possible paths are followed when one observes actual performances of the task, but only a small proportion of them. In most of the analysis performed on microworld data the experimenter has to restrict the variability by imposing some structure (a-priori, theoretically driven assumptions) on the data. For example, hypothetical ways of action –strategies- can be described and implemented, and participants’ performances can be described in terms of their similarity to the strategies. However, the selection of this theoretical structure (How many strategies are possible? How many are representative enough? Are they generalizable to different conditions?) is indefectibly biasing the results obtained in the analysis. In that sense, the LSA approach is self-organizing, and does not require defining an a –priori theoretical structure.
Before we start describing what LSA is and how it can be applied to CPS, we would like to stress some abstract considerations that underlie the approach that we are about to implement. These considerations are independent of the procedure itself (other procedures could be defined using this framework), but, in our opinion, an essential step to dealing with the complexity of the tasks at hand: (1) Each microworld can be conceptualized as a complex, multidimensional feature space. (2) To address the intractability problem, we usually need to create a representation or transformation of this original multidimensional feature space. To do this, we need to find a set of features that represent the characteristics that make participants different, and to obviate those that are not important. (3) Last, each trial of every subject can be conceptualized as an implementation of several values in the feature space. Not only a trial, but every subpart or superpart of a participant performance (strategies or performance patterns) can be thought of as an object in this space.
LSA is one implementation of this framework, and in the following parts of this paper we try to show why we think that it is a successful one. In the next section, we briefly describe Firechief, the microworld that we use for our examples. An introduction to LSA follows, explaining what LSA is and how it has been used as a theory of knowledge representation and as a tool for text-comprehension-related applications. Next we explain its procedure and mathematical foundations. After that, we describe how LSA can be applied to microworlds and show some examples of use. We then compare LSA similarities to judgments of relatedness emitted by human participants exposed to similar information to validate the technique. As a conclusion, we discuss the importance of the approach and the possibilities and future extensions of the technique.
1. A description of the example application task
In our examples, we use the microworld Firechief [27], which simulates a forest where a fire is spreading. Their task is to extinguish the fire as soon as possible. In order to do so, they can use helicopters and trucks (each one with particular characteristics) that can be controlled by mouse movements and key presses. There
are four commands that are used to control the movement and functions of the appliances: (1) Drop water on the current landscape segment; (2) Start a control fire (trucks only) on the current landscape segment; (3) Move an appliance to a specified landscape segment, and (4) Design an automatic surveillance area where the vehicle will find and extinguish fires automatically. Command 4 is not considered in the following discussion and was disabled, so it is not further described. Firechief generates pretty complex log files containing every action a participant has done. We have been analyzing these actions to classify participants, in order to detect which strategies lead good performers to a bad performance when the environmental situation changes slightly [28], [5].
Every time a participant perform an action, it is saved in a log file as a row containing action number, command (e.g. drop water or move) or event2 (e.g., a wind change or a new fire), current performance score, appliance number, appliance type, position, and landscape type. Most of these variables are not continuous, but on a nominal scale, such as type of movement. For more information on the structure of the log files, see [27].
The set of trials that was used in this report (referred as corpus) is formed by four experiments (described in Quesada, et al. [28] and Ca?as et al. [5]). The particular hypothesis, manipulations and results are not of interest for our current purpose of explaining the application of LSA to microworld log files.
2. Introduction to LSA
LSA is a machine-learning model that induces representations of the meaning of words by analyzing the relation between words and passages in large bodies of representative text. LSA is both a method (tool) used in industry to develop technology to improve educational applications, and a theory of knowledge representation used to model well known experimental effects in text comprehension and priming, among others [24]. Latent Semantic Analysis was originally developed in the context of information retrieval [6] as a way of overcoming problems with polysemy and synonymy. Some words appear in the same contexts (synonyms) and an important part of word usage patterns is blurred by accidental and inessential information. As Kintsch [19] put it, why did an author choose a particular word in a specific place instead of some other alternative? The method used by LSA to capture the essential semantic information is dimension reduction, selecting the most important dimensions from a co-occurrence matrix decomposed using Singular Value Decomposition (see below). As a result, LSA offers a way of assessing semantic similarity between any two samples of text in an automatic, unsupervised way.
2 Events are generated by the system, while actions are generated by the user. Events are also lines in the log file. Only 1-2% of the lines in a log file are events.
LSA has been used in applied settings with a surprising degree of success in areas like automatic essay grading [9] and automatic tutoring to improve summarization skills in children [19]. As a model, LSA’s most impressive achievements have been in human language acquisition simulations [24] and in modeling of high-level comprehension phenomena like metaphor understanding, causal inferences and judgments of similarity [20].
The best way of understanding the LSA induction mechanism is through an example, on text passages and words, and the best small-scale example is described in Landauer, Foltz and Laham [25]3. The reader is strongly recommended to consult these references for a better understanding of the technique.
Although LSA has been mostly used on text corpora, our basic point is that LSA can be applied to any domain of knowledge where there are a high number of weak relations between tokens, as in Complex Problem Solving log files. Instead of using word co-occurrence statistics and huge samples of text, we have used a representative amount of activity in controlling dynamic systems, and actions or states have been used to develop the much-wanted objective measure of similarity in the changing, time-dependent, highly complex experimental tasks known as microworlds. The next sections show the basic steps to perform the analysis and will present some examples of the powerful analysis that can be conducted.
3. LSA procedure
The procedure starts with creating a matrix of actions4 by trials. If the log files record state information instead of actions, states can be used5. Note that this is not an exhaustive state space, or a mapping of all possible transitions between actions (since in most of the systems –other than small ones like Hayes and Broadbent’s sugar factory and the like- this task would be excessively demanding, see Buchner, Funke and Berry, [3]). Our corpus was composed of 360,199 actions in 3441 trials. Among them, only 75,565 were different actions, which means that on average each action appears 6.25 times in the corpus. Note that we are representing only the information that actual people interacting with the system experienced, not all possible actions in this microworld.
3 The same example has been used also in Deerwester et al. [6] and Landauer and Dumais [24].
4 The information contained in each action (in this microworld, an action is one line of code in the log files) was transformed into a single token using underscores instead of the spaces that separated values of different variables.
5 From here on, we talk about actions as the basic token. However, states can be considered the basic token with no further assumptions for the model. We use actions because Firechief states are forest matrices (24x15) containing information about type of terrain and fire status (safe, burned, fire intensity). In this particular example it is more convenient to use actions than states, although in other microworlds like DuressII states are a better option (see Quesada, Kintsch, and Gomez, this issue).
Each of these 75,565 rows stands for a unique action, and each of the 3441 columns stand for a trial. Each cell contains the frequency with which the action of its row appears in the trial denoted by its column. Note that most of the cells will contain a frequency of zero, since most actions appear in only a few trials and do not appear in the rest.
This matrix of frequencies is decomposed using Singular Value Decomposition (SVD). In SVD, a rectangular matrix is decomposed into the product of three other matrices. The first matrix is containing the original row entities but the columns are derived orthogonal factor values; the third matrix describes the original column entities in the same manner; the second (and most important) matrix is a diagonal matrix containing scaling values such that when the three matrices are multiplied, the result is the original matrix. It has been mathematically proven that any matrix can be decomposed and then recomposed perfectly using only as many factors as the smallest dimension of the original matrix. However, the interesting phenomenon occurs when the original matrix is recomposed using fewer dimensions than necessary: the reconstructed matrix is a least-squares best fit. This feature has been successfully used in fields ranging from satellite information compression to solving huge linear equation systems (e.g., [32]).
When the actions-by-trials matrix is recomposed using a small fraction of the available dimensions (usually between 100 and 300 first dimensions, zeroing out the rest), the new matrix contains information that has been inferred from the dependencies between actions and the context where these actions appeared. In fact, the contexts where these actions did not appear are as important -carry as much information- as those where they did. The microworld is a new multidimensional feature space, where both actions and context (trials) are represented in a way that amplifies those characteristics that make participants different, and obviate those that are not important for classifying their performance. Other features of the theory, such as the election of the cosine between vectors as a measure of similarity6, or the weighting schemes that are applied to the frequency matrix7 are of less importance.
6 Cosine and correlation have a very similar interpretation. But classic studies on information retrieval [18] address an important advantage in the cosine measure. Both measures can be conceptualized as an inner product divided by the product of vector lengths. The main difference is that in correlation the scores are normalized by subtracting the mean of each dimension. This normalization creates a vector whose components are deviations of corresponding components in the original vector from the mean of the original vector. This causes the correlation measure to lose angle monotonicity and radial monotonicity, and it is why the cosine measure is preferred.
7 See [15].
4. LSA Applied to Microworlds: S ome Examples and Possible Analysis
This transformation, and its particular use in former applications of LSA to infer word meaning, can be related in a very interesting way when the application to Complex Problem Solving is considered. Some actions can be considered as functional synonyms: they appear in the same contexts, and fulfill approximately the same function. The following example illustrates this idea.
In Table 1, some actions are defined. For simplicity, some variables that are normally contained in the log files have been removed8. Example1 contains a movement to the point (11, 9) in the screen, which is of type forest, and then, a drop water action there. Example 3 shows a very similar picture, where the movement is done to a contiguous cell (10,9) that is also of type forest. From a human point of view, these two examples are highly similar. For LSA they are too, as can be seen in their similarity expressed as a cosine of 0.854 in Table 2.
The second example has a rather different ‘meaning’ since the cell targeted is (15,15), quite far from the cell used in examples 1 and 3. The cosines between them and example2 (.124 and .125) are, accordingly, smaller than the one between 1 and 3. Example 4 describes an action that has been performed in the same cell as in example 1 (11,9), but this time is a control fire instead of a drop-water action. The cosine between 1 and 4 is high (0.56), expressing a certain similarity between the two actions, but not as high as in examples 1 and 3, where the objective similarity is more evident.
Tables 3 and 4 depict a more complex example where wider slices of performance (8 actions) are compared. The samples labeled example1, example2, and example 3 are beginnings of trials that have been selected randomly from the corpus. This time, all the usable information contained in the log file is displayed. Each action has six components: type of action, appliance number, appliance type, departure cell, arrival cell and type of arrival cell.
Time t1Time t2
Example 1 move_11_9_forest drop_11_9_forest
Example 2 move_15_15_forest drop_15_15_forest
Example 3 move_10_9_forest drop_10_9_forest
Example 4 move_11_9_forest control_11_9_forest
Table 1: Example of how LSA captures similarity at a very molecular level (action level). Time T1 and Time T2 are two consecutive snapshots of performance, containing one action each.
8 Types of appliance, appliance number and departure point have been removed, as well as generation number (a timestamp), and overall performance score at this particular moment. See Omodei and Wearing [27] for more information on the variables.
Example 1Example 2Example 3Example 4 Example 10.1240490.8543280.6626815
Example 20.12596550.0772655
Example 30.5660885
Example 4
Table 2: Comparisons between the four examples. Cells are cosines that represent similarity between all possible pairs. These have been obtained by averaging the cosines for different temporal moments. This means that the value in cell example1 – example2 has been calculated as the average of the cosine (example1 in t1 vs. example2 in t1) and the cosine (example1 in t2 vs. example2 in t2 ).
One difficulty arises. When LSA is used on text, cosines are easily understood since every reader has an intuitive experience of meaning (e.g., the sentences ‘The man was driving a yellow car’ and ‘The man was traveling in a red car’ have a cosine of .89, and our common sense tells us that these sentences convey similar information). When LSA is used on samples of performance from a microworld, there is no way the reader can understand the ‘meaning’ of the log files without watching a replay or having an extraordinarily vivid imagination plus experience with the task. For most researchers, the following extracts in table 3 are hardly understandable. For researchers familiar with Firechief, they should be as clear as a piece of sheet music to a musician. However, understanding the contents of these examples is not conditio sine qua non for understanding the advantage of LSA analysis over two other methods, namely exact matching and correlation between transition matrices. Suffice it to say that examples 1 and 2 are very similar and example 3 is very different from them. The attentive observer could induce this from the locations (coordinates in the Firechief map), the type of actions, and type of landscape cell.
An exact matching method would count the number of times that the same action is in two examples. Then, the number of matchings divided by the total number of actions in the example would be the similarity between two samples. This method would render a similarity of 1/8 between example 1 and 2, and zero in comparisons 1 vs. 3 and 2 vs. 3. This method is an equivalent to keyword counting in text and has been proven to be insufficient to capture similarity in meaning, because of the polysemy and synonymy effects described before.
A somewhat more flexible method is the use of transitions between actions, proposed by Howie and Vicente [17] and used in Quesada et al. [28] and Ca?as et al. [5]. It is based on counting the number of times that one type of action precedes any other type. The frequencies of every transition are registered in cells in a table, and then the resulting tables for two examples are correlated. The method cannot account for all the variability in actions, because of the huge amount of zero entries that artificially increase the correlation, so only action type was considered. This analysis is shown in tables 4(a,b,c). Since lots of information contained in the log files has been dropped,
the method does not distinguish between these examples. The correlation between table a and b is 0.971; exactly the same correlation is obtained for tables b and c, and the comparison between a and c is 1 since the sequence of type of action is exactly the same. Note that even though this method is seriously limited and is showing very flawed similarity estimations, it has been used in several works in the literature.
example1Example2
move_2_truck_4_11_13_3_forest move_2_truck_4_11_12_15_forest
move_1_truck_4_14_16_14_forest move_1_truck_4_14_13_5_forest
move_3_copter_8_6_11_12_forest move_4_copter_11_4_11_9_forest
move_4_copter_11_4_11_9_forest drop_water_4_copter_11_9_forest
control_fire_2_truck_13_3_forest move_4_copter_11_9_13_8_forest
control_fire_1_truck_16_14_forest control_fire_2_truck_12_15_forest
move_2_truck_13_3_17_7_clearing move_2_truck_12_15_13_14_forest
move_1_truck_16_14_20_12_forest control_fire_2_truck_13_14_forest
example1Example3
move_2_truck_4_11_13_3_forest move_2_truck_4_11_2_2_pasture
move_1_truck_4_14_16_14_forest move_1_truck_4_14_0_5_forest
move_3_copter_8_6_11_12_forest move_4_copter_8_6_8_4_clearing
move_4_copter_11_4_11_9_forest move_3_copter_8_6_8_10_clearing
control_fire_2_truck_13_3_forest control_fire_2_truck_2_2_pasture
control_fire_1_truck_16_14_forest control_fire_1_truck_0_5_forest
move_2_truck_13_3_17_7_clearing move_4_copter_8_4_4_2_forest
move_1_truck_16_14_20_12_forest move_3_copter_8_10_2_3_clearing
Table 3: First 8 movements in 3 slices randomly sampled from the Firechief experiments described in Quesada et al., [27] and Ca?as et al. [5]. When an action is shared by two extracts, it is marked as a shaded cell.
Finally, let us look at the results of similarity estimation using LSA cosines. The vector representing the sample has been calculated as the average of the 8 action vectors. Example 1 vs. example 2 has a cosine of 0.721928, a high similarity value. Even though these samples share only 1/8 of the actions, LSA has correctly inferred that the remaining actions, although different, are functionally related. Comparisons between 1-3 and 2-3 have cosines as low as 0.056770 and 0.071135 respectively, showing that these performances are different indeed.
Table 5 shows a comparative between the three methods. The exact matching method is characterized by an underestimation of the similarity. Since the current examples only have 8 actions from the beginning of the trial and one happened to be the same in examples 1 and 2, the similarity estimated by this method is not zero, but .125. However, in real examples where the number of actions is much higher, the
probability of finding repeated actions decreased, and so does the similarity calculated with this method, no matter what the objective similarity is. This method would predict correctly the lack of similarity, but it is not sensitive to real similarity. Since virtually every pair of trials in Firechief shares very few actions, or even none, even when they are objectively similar, this method is not appropriate.
(a)
(b)Example 1
Example 2
drop move control drop move control drop
000drop 010move
041move 122control 0
11control 010(c)
Example 3
drop move control drop
000move
041control 011Table 4 Transitions between actions considering type of action only as described in Quesada et al., [27] and Ca?as et al. [5], for the examples 1,2 and 3. Cells contain frequencies of the transition defined by its row and its column. For instance, the number 4 in the center cell in table 4a means that in example 1 the transition move-move has appeared four https://www.doczj.com/doc/4517646641.html,parisons method
example1 to example2example1 to example3example2 to example3Objective similarity
high low Low LSA Cosine
0.72190.05670.0711Exact matching
0.12500Transitions between actions 0.97110.971Table 5 Similarity between examples 1, 2, and 3 using the methods LSA, exact matching and transitions between actions as described in Quesada et al., [27] and Ca?as et al. [5].
On the other hand, the Transitions between actions method is characterized by an overestimation of the similarity. Most of the information contained in the log files is disregarded due to the limitations of the technique (exponential growing of the number of possible pairs). This produces that series of actions that are completely unrelated are considered exactly the same because the ‘type of action’ information is coincidentally the same, as in examples 1 and 3. In contrast, the similarity values computed by LSA reflect the objective similarity very closely.
5. Correlations between LSA and Human Judgment
An important point is that LSA similarity judgments can be compared with human standards. Up to this point in the paper, only the experimenter’s common sense and knowledge of the task have been used to validate the results. However, if LSA captures similarity between complex problem solving performances in a meaningful way, any person with experience on the task could be used as a validation. The problem is that, contrary to what happens when one uses LSA to model text comprehension, it is not easy to find experts in the task at hand. Everybody is a perfect example of the expert reader, but not everybody is an expert in controlling the particular dynamic system called Firechief. To test our assertions about LSA, we recruited 3 persons and exposed them to the same amount of practice as our experimental participants, so they could learn the constraints of the task.
After 24 practice trials, these participants were used to assess the external validity of LSA similarities. Using Firechief’s replay option, participants had to watch 7 pairs of trials (at a pace faster than normal) and express similarity judgments about these pairs. People watched a randomly ordered series of trials, in a different order for each participant, which were selected as a function of the LSA cosines (pairs A, B, C, D, E, F, G with cosines 0.75, 0.90, 0.53, 0.60, 0.12 and 0.06 respectively). One of the pairs was presented twice to measure test-retest reliability. That is, for example, pair G was exactly the same as pair A for one participant, the same as pair F for another participant, etc. Filling out a form that presented all the possible pairings of ‘stimuli pairs’9 were presented. They had to answer which pair seemed more similar to them. For example, LSA would say that pair B is more related (>) than pair C, since the cosines are 0.90 and .53 respectively.
LSA cosines predicted human similarity judgments very well indeed. For 3 participants in this pilot study, the proportion of agreement LSA-human was 6/19, 14/19, and 13/19 respectively. Participants with strong agreement with LSA also showed more consistency in their judgments, that is they answered to the repeated item in the same way. The participant who had low agreement with LSA had a very bad performance grading the repeated item, which suggests that she could have been answering randomly. Even so, the average agreement between LSA cosines and human judgments was 0.57, far superior to the agreement expected by chance, 0.5 ^ 19 = 2e-5.
However, some problems have to be considered. The application of any analysis based on contingency tables was not possible due to two issues. First, because of the small number of observations (19) some cells in the contingency table had a frequency less than five, which make the results of procedures such as the phi coefficient unreliable. Second, the stimuli were presented using all possible combinations, that is AB, AC, AD, … FG. Due to the characteristics of the stimuli 9 This method was selected so the cosines were very distinctive and easy to compare. A subspace where a few items (instead of pairs) have very distinctive cosines could be found too, but this approach was used because the sample of stimuli was easier to obtain.
(see cosine list), most of the pairs should be answered with a ‘greater than’ response, and this is contrary to the assumptions of the phi coefficient, that will give flawed results if there is an extreme imbalance in the marginal distributions of the contingency table. As a consequence of this flaw in the design of the human judgments pilot experiment we cannot distinguish clearly if the agreement between LSA and participant was due to a response bias ‘answer always greater than’ or to the fact that the similarity that participants perceived was captured by the LSA cosines. Future designs will address these shortcomings, for example having a greater number of items to judge, and presenting them in a way that distributes the correct responses equally (example half of the options in the form BA, and half AB). A possible alternative is to substitute the dichotomy in the answer with an ordinal, or even interval scale, but we still need to make sure that participants are able to do fine-grained distinctions when emitting similarity judgments for replays of this kind of tasks.
5. Conclusions
LSA seems to be a promising new way of approaching Complex Problem Solving performance that overcomes some of the known limitations of previous methods. Apart from the features listed in the introduction, there are some pragmatic LSA advantages worth noting: (1) Since the basic unit of analysis is the token (action or state), even systems that are described in terms of nominal (discrete) variables can be analyzed. Both actions and states can be used as units. (2) The ‘semantic’ matching mechanism permits discovery of similarities beyond simple coincidence in the log files containing actions or states. That is, participants who are using different interventions to realize the same strategy will be considered similar even if their log files share no actions (or states). (3) The level of granularity (whether we are working with individual tokens, slices of performance, whole trials, or collections of trials) is not defined a-priori. Since every object, from one token to the participant’s whole performance, can be represented as a vector of n dimensions, analysis can be performed at any level of detail.
Not less important are the following disadvantages: (1) A huge sample of data is needed. (2) Order effects are not taken into account. This means that for LSA a trial where the tokens have been scrambled to a random order has exactly the same meaning as the original version. This could be considered an important shortcoming (although see [26] for the contrary view), and current research is being performed in this direction. (3) Though the SVD analysis is common practice and can be found in several statistical packages10, a powerful computer is needed to run large analysis. One more question is in the air: Is LSA a method or a theory of CPS? The answer may be: both. In a seminal paper, Gigerenzer [13] argued that the way social scientists 10 R and matlab have an SVD function. To inquire about LSA computer programs, address Telcordia (Formerly know as Bellcore, who owns the patent for information retrieval) at https://www.doczj.com/doc/4517646641.html,/
discover theories of mind is closely bound to the emergence of new tools for data analysis, rather than new data. He exemplifies this phenomenon with the signal detection theory, a cognitive theory of perception that is a generalization of the Neyman-Pearsonian statistical hypothesis testing technique. The tools-to-theory proposal exemplifies perfectly the situation of LSA as a statistical method that can be understood as a theory of knowledge representation. At this moment, LSA is a tool that can be used to complement and extend the reach of some other non-computational theories. However, learning in a CPS environment is often thought of as predicting each successive constituent from those already analyzed on the basis of the knowledge acquired from past experience with the system (e.g. [11]). It should not be too difficult to employ LSA to implement that proposal. The LSA approach to this idea would rely on the following assumptions: (a) Actions would be generated by comparing a small amount of preceding context with a vast amount of knowledge acquired in past interactions with the system. (b) The adequacy of an action would be computed by examining past contexts where this action has appeared, and comparing this set of contexts with the current one. This assumption would agree with recognition-based theories such as Klein’s Recognition Primed Decision Making ([21], [22]). (c) In LSA, context and actions (or states) are both represented as vectors, and similarity is estimated using the cosine between the vectors. In normal conditions (lots of knowledge about the situation), the action selected would be the one with the highest cosine with the current context. However, the validity of these assumptions is not assured, and future research should address this issue. Acknowledgements
Our acknowledgements to Tom Landauer for proposing interesting issues concerning the selection of the unit of analysis in Complex Problem Solving. We are grateful to Kim Vicente and John Hajdukiewicz for sharing experimental data and insightful discussions during a visit of José Quesada to the University of Toronto. Many thanks to Bill Oliver, who provided passionate methodological discussions and theoretical contributions. The manuscript was thoroughly revised by Nancy Mann, improving greatly the language and presentation.
This research was in part supported by Grant EIA – 0121201 from the national science foundation.
References
1. Brehmer, B., Feedback delays in complex dynamic decision tasks. In P. Frensch and J.
Funke, (Eds.) Complex Problem Solving: The European Perspective. Hillsdale, NJ: Lawrence Erlbaum (1995)
2. Broadbent, D. E. Levels, hierarchies and the locus of control. Quarterly Journal of
Experimental Psychology 32, (1977) 109-118
3. Buchner, A., Funke, J., Berry, D.: Negative correlations between control performance and
verbalizable knowledge: Indicators for implicit learning in process control tasks? Quarterly Journal of Experimental Psychology, 48A, (1995) 166-187
4. Calderwood, R.: The role of context in modeling domain knowledge. Unpublished doctoral
dissertation, University of New Mexico (1989)
5. Ca?as, J.J., Quesada, J.F., Antolí, A., Fajardo, I.: Cognitive flexibility and adaptability to
environmental changes in dynamic complex problem solving tasks (submitted to ergonomics)
6. Deerwester, S., Dumais, S. T., Furnas, G. W., Landauer, T. K., Harshman, R. Indexing By
Latent Semantic Analysis. Journal of the American Society For Information Science, 41, (1990) 391-407
7. Dickson, J., McLennan, J, Omodei, M. M.: Effects of concurrent verbalization on a time
pressured dynamic decision task. Journal of General Psychology. 127, (2000) 217-228
8. D?rner, D.: Wie Menschen eine Welt verbessern wollten und sie dabei zerst?rten How
people wanted to improve the world. Bild der Wissenschaft, Heft 2 (popul?rwissenschaftlicher Aufsatz) (1975)
9. Foltz, P. W. , Laham, D., Landauer, T. K.,: The Intelligent Essay Assessor: Applications to
Educational Technology Interactive Multimedia Education Journal of Computer enhanced learning On-line journal., 1(2). https://www.doczj.com/doc/4517646641.html,/articles/1999/2/04/index.asp (1999) 10. Frensch, P., Funke, J.: Complex Problem Solving: The European Perspective. Hillsdale, NJ:
Lawrence Erlbaum (1995)
11. Funke, J.: Dealing with dynamic systems: Research strategy, diagnostic approach and
experimental results. German Journal of Psychology, 16, (1992) 24-43
12. Gibson, F. P.: Feedback delays: How can decision makers learn not to buy a new car every
time the garage is empty? Organizational Behavior and Human Decision Processes Vol. 83, No. 1 (2000) 141 - 166
13. Gigerenzer, G.: From tools to theories: A heuristic of discovery in cognitive psychology.
Psychological Review, 98 (1991) 254-267
14. Green D. W.: Understanding microworlds. Quarterly Journal of Experimental Psychology,
Section A-Human Experimental Psychology, 54 (3) (2001) 879-901
15. Harman, D. An experimental study of the factors important in document ranking. In
Association for Computing Machinery Conference on Research and Development in Information Retrieval, New York: Association for Computing Machinery (1986)
16. Henderson, L., Klemes, J., Eshet, Y.: Just playing a game? Educational simulation software
and cognitive outcomes. Journal of Educational Computing Research, 22 (1) (2000) 105-129
17. Howie, D. E., Vicente, K. J.: Measures of operator performance in complex, dynamic
microworlds: Advancing the state of the art, Ergonomics, vol. 41, (1998) 85-500
18. Jones W. P. and Furnas, G. W.: Pictures of relevance: A Geometric Analysis of similarity
measures. Journal of the American society for information science, 38(6) (1987) 420 -442 19. Kintsch, E., Steinhart, D., Stahl, G., Matthews, C., Lamb, R., the LSA Research Group
Developing summarization skills through the use of LSA-backed feedback, Interactive Learning Environments, 8 (2), (2000) 87-109
20. Kintsch, W. Predication. Cognitive Science 25, (2001) 173-202
21. Klein G.: A Recognition-primed decision model of rapid decision making. In Klein, G.A.,
Orasanu, J., Calderwood, R., Zsambok, C.E. (Eds.), Decision making in action: Models and methods. Norwood, NJ: Ablex Publishing Corporation, (1993)
22. Klein, G.: The recognition-primed decision model: looking back, looking forward. In
Zsambok C., Klein, G. (Eds.), Naturalistic Decision Making. Mahwah, N.J.: Lawrence Erlbaum Associates, 1997
23. Klein, G.A., Orasanu, J., Calderwood, R., Zsambok, C.E. (Eds.), Decision
making in action: Models and methods. Norwood, NJ: Ablex Publishing Corporation, (1993)
24. Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato's problem: The Latent
Semantic Analysis theory of the acquisition, induction, and representation of knowledge.
Psychological Review, 104, 211-240
25. Landauer, T. K., Foltz, P. W., Laham, D. (1998). Introduction to Latent Semantic Analysis.
Discourse Processes, 25, 259-284
26. Landauer, T. K., Laham, D., Rehder, B., Schreiner, M. E.,: How well can passage meaning
be derived without using word order? A comparison of Latent Semantic Analysis and humans. In Shafto, M. G., Langley, P. (Eds.) Proceedings of the 19th annual meeting of the Cognitive Science Society. Mahwah, NJ: Erlbaum (1997) 412-417
27. Omodei, M. M., & Wearing, A. J. (1995). The Fire Chief microworld generating program:
An illustration of computer-simulated microworlds as an experimental paradigm for studying complex decision-making behavior. Behavior Research Methods, Instruments & Computers, 27, 303-316
28. Quesada, J.F., Ca?as, J.J., Antoli, A.: An explanation of human errors based on
environmental changes and problem solving strategies. In Wright, P., Dekker, S., Warren
C.P. (Eds.) ECCE-10: Confronting Reality. Sweden: EACE (2000)
29. Rasmussen, J.: The role of hierarchical knowledge representation in decision making and
system management. IEEE Transactions on Systems, Man, and Cybernetics, SMC-15 (2): (1985) 234-243
30. Shepard, R.N. (1987). Toward a universal law of generalization for psychological science.
Science, 237, 1317-1323
31. Simon, H.A.: The sciences of the artificial. MIT press (1981)
32. Strang, G.: Linear Algebra and Its Applications. Harcourt, Brace and Jovanovich (1988)
33. Vicente, K.: Cognitive Work Analysis: Toward Safe, Productive, and Healthy Computer-
based work. Lawrence Erlbaum associates, London (1999)
34. Winograd, T: A procedural model of language understanding. In Schank, R., Colby, K.
(eds.) : Computers models of thought and Language, Sand Francisco W.H. Freeman (1972)
从实践的角度探讨在日语教学中多媒体课件的应用 在今天中国的许多大学，为适应现代化，信息化的要求，建立了设备完善的适应多媒体教学的教室。许多学科的研究者及现场教员也积极致力于多媒体软件的开发和利用。在大学日语专业的教学工作中，教科书、磁带、粉笔为主流的传统教学方式差不多悄然向先进的教学手段而进展。 一、多媒体课件和精品课程的进展现状 然而，目前在专业日语教学中能够利用的教学软件并不多见。比如在中国大学日语的专业、第二外語用教科书常见的有《新编日语》（上海外语教育出版社）、《中日交流标准日本語》（初级、中级）（人民教育出版社）、《新编基础日语（初級、高級）》（上海译文出版社）、《大学日本语》（四川大学出版社）、《初级日语》《中级日语》（北京大学出版社）、《新世纪大学日语》（外语教学与研究出版社）、《综合日语》（北京大学出版社）、《新编日语教程》（华东理工大学出版社）《新编初级（中级）日本语》（吉林教育出版社）、《新大学日本语》（大连理工大学出版社）、《新大学日语》（高等教育出版社）、《现代日本语》（上海外语教育出版社）、《基础日语》（复旦大学出版社）等等。配套教材以录音磁带、教学参考、习题集为主。只有《中日交流標準日本語（初級上）》、《初級日语》、《新编日语教程》等少数教科书配备了多媒体DVD视听教材。 然而这些试听教材，有的内容为日语普及读物，并不适合专业外语课堂教学。比如《新版中日交流标准日本语（初级上）》，有的尽管DVD视听教材中有丰富的动画画面和语音练习。然而，课堂操作则花费时刻长，不利于教师重点指导，更加适合学生的课余练习。比如北京大学的《初级日语》等。在这种情形下，许多大学的日语专业致力于教材的自主开发。 其中，有些大学的还推出精品课程，取得了专门大成绩。比如天津外国语学院的《新编日语》多媒体精品课程为2007年被评为“国家级精品课”。目前已被南开大学外国语学院、成都理工大学日语系等全国40余所大学推广使用。
Unit1 Americans believe no one stands still. If you are not moving ahead, you are falling behind. This attitude results in a nation of people committed to researching, experimenting and exploring. Time is one of the two elements that Americans save carefully, the other being labor. "We are slaves to nothing but the clock,” it has been said. Time is treated as if it were something almost real. We budget it, save it, waste it, steal it, kill it, cut it, account for it; we also charge for it. It is a precious resource. Many people have a rather acute sense of the shortness of each lifetime. Once the sands have run out of a person’s hourglass, they cannot be replaced. We want every minute to count. A foreigner’s first impression of the U.S. is li kely to be that everyone is in a rush -- often under pressure. City people always appear to be hurrying to get where they are going, restlessly seeking attention in a store, or elbowing others as they try to complete their shopping. Racing through daytime meals is part of the pace
新视野三版读写B2U4Text A College sweethearts 1I smile at my two lovely daughters and they seem so much more mature than we,their parents,when we were college sweethearts.Linda,who's21,had a boyfriend in her freshman year she thought she would marry,but they're not together anymore.Melissa,who's19,hasn't had a steady boyfriend yet.My daughters wonder when they will meet"The One",their great love.They think their father and I had a classic fairy-tale romance heading for marriage from the outset.Perhaps,they're right but it didn't seem so at the time.In a way, love just happens when you least expect it.Who would have thought that Butch and I would end up getting married to each other?He became my boyfriend because of my shallow agenda:I wanted a cute boyfriend! 2We met through my college roommate at the university cafeteria.That fateful night,I was merely curious,but for him I think it was love at first sight."You have beautiful eyes",he said as he gazed at my face.He kept staring at me all night long.I really wasn't that interested for two reasons.First,he looked like he was a really wild boy,maybe even dangerous.Second,although he was very cute,he seemed a little weird. 3Riding on his bicycle,he'd ride past my dorm as if"by accident"and pretend to be surprised to see me.I liked the attention but was cautious about his wild,dynamic personality.He had a charming way with words which would charm any girl.Fear came over me when I started to fall in love.His exciting"bad boy image"was just too tempting to resist.What was it that attracted me?I always had an excellent reputation.My concentration was solely on my studies to get superior grades.But for what?College is supposed to be a time of great learning and also some fun.I had nearly achieved a great education,and graduation was just one semester away.But I hadn't had any fun;my life was stale with no component of fun!I needed a boyfriend.Not just any boyfriend.He had to be cute.My goal that semester became: Be ambitious and grab the cutest boyfriend I can find. 4I worried what he'd think of me.True,we lived in a time when a dramatic shift in sexual attitudes was taking place,but I was a traditional girl who wasn't ready for the new ways that seemed common on campus.Butch looked superb!I was not immune to his personality,but I was scared.The night when he announced to the world that I was his girlfriend,I went along
Unit 1 1学习外语是我一生中最艰苦也是最有意义的经历之一。虽然时常遭遇挫折，但却非常有价值。 2我学外语的经历始于初中的第一堂英语课。老师很慈祥耐心，时常表扬学生。由于这种积极的教学方法，我踊跃回答各种问题，从不怕答错。两年中，我的成绩一直名列前茅。 3到了高中后，我渴望继续学习英语。然而，高中时的经历与以前大不相同。以前，老师对所有的学生都很耐心，而新老师则总是惩罚答错的学生。每当有谁回答错了，她就会用长教鞭指着我们，上下挥舞大喊：“错！错！错！”没有多久，我便不再渴望回答问题了。我不仅失去了回答问题的乐趣，而且根本就不想再用英语说半个字。 4好在这种情况没持续多久。到了大学，我了解到所有学生必须上英语课。与高中老师不。大学英语老师非常耐心和蔼，而且从来不带教鞭！不过情况却远不尽如人意。由于班大，每堂课能轮到我回答的问题寥寥无几。上了几周课后，我还发现许多同学的英语说得比我要好得多。我开始产生一种畏惧感。虽然原因与高中时不同，但我却又一次不敢开口了。看来我的英语水平要永远停步不前了。 5直到几年后我有机会参加远程英语课程，情况才有所改善。这种课程的媒介是一台电脑、一条电话线和一个调制解调器。我很快配齐了必要的设备并跟一个朋友学会了电脑操作技术，于是我每周用5到7天在网上的虚拟课堂里学习英语。 6网上学习并不比普通的课堂学习容易。它需要花许多的时间，需要学习者专心自律，以跟上课程进度。我尽力达到课程的最低要求，并按时完成作业。 7我随时随地都在学习。不管去哪里，我都随身携带一本袖珍字典和笔记本，笔记本上记着我遇到的生词。我学习中出过许多错，有时是令人尴尬的错误。有时我会因挫折而哭泣，有时甚至想放弃。但我从未因别的同学英语说得比我快而感到畏惧，因为在电脑屏幕上作出回答之前，我可以根据自己的需要花时间去琢磨自己的想法。突然有一天我发现自己什么都懂了，更重要的是，我说起英语来灵活自如。尽管我还是常常出错，还有很多东西要学，但我已尝到了刻苦学习的甜头。 8学习外语对我来说是非常艰辛的经历，但它又无比珍贵。它不仅使我懂得了艰苦努力的意义，而且让我了解了不同的文化，让我以一种全新的思维去看待事物。学习一门外语最令人兴奋的收获是我能与更多的人交流。与人交谈是我最喜欢的一项活动，新的语言使我能与陌生人交往，参与他们的谈话，并建立新的难以忘怀的友谊。由于我已能说英语，别人讲英语时我不再茫然不解了。我能够参与其中，并结交朋友。我能与人交流，并能够弥合我所说的语言和所处的文化与他们的语言和文化之间的鸿沟。 III. 1. rewarding 2. communicate 3. access 4. embarrassing 5. positive 6. commitment 7. virtual 8. benefits 9. minimum 10. opportunities IV. 1. up 2. into 3. from 4. with 5. to 6. up 7. of 8. in 9. for 10.with V. 1.G 2.B 3.E 4.I 5.H 6.K 7.M 8.O 9.F 10.C Sentence Structure VI. 1. Universities in the east are better equipped, while those in the west are relatively poor. 2. Allan Clark kept talking the price up, while Wilkinson kept knocking it down. 3. The husband spent all his money drinking, while his wife saved all hers for the family. 4. Some guests spoke pleasantly and behaved politely, while others wee insulting and impolite. 5. Outwardly Sara was friendly towards all those concerned, while inwardly she was angry. VII. 1. Not only did Mr. Smith learn the Chinese language, but he also bridged the gap between his culture and ours. 2. Not only did we learn the technology through the online course, but we also learned to communicate with friends in English. 3. Not only did we lose all our money, but we also came close to losing our lives.
新大学日语简明教程课文翻译 第21课 一、我的留学生活 我从去年12月开始学习日语。已经3个月了。每天大约学30个新单词。每天学15个左右的新汉字，但总记不住。假名已经基本记住了。 简单的会话还可以，但较难的还说不了。还不能用日语发表自己的意见。既不能很好地回答老师的提问，也看不懂日语的文章。短小、简单的信写得了，但长的信写不了。 来日本不久就迎来了新年。新年时，日本的少女们穿着美丽的和服，看上去就像新娘。非常冷的时候，还是有女孩子穿着裙子和袜子走在大街上。 我在日本的第一个新年过得很愉快，因此很开心。 现在学习忙，没什么时间玩，但周末常常运动，或骑车去公园玩。有时也邀朋友一起去。虽然我有国际驾照，但没钱，买不起车。没办法，需要的时候就向朋友借车。有几个朋友愿意借车给我。 二、一个房间变成三个 从前一直认为睡在褥子上的是日本人，美国人都睡床铺，可是听说近来纽约等大都市的年轻人不睡床铺，而是睡在褥子上，是不是突然讨厌起床铺了？ 日本人自古以来就睡在褥子上，那自有它的原因。人们都说日本人的房子小，从前，很少有人在自己的房间，一家人住在一个小房间里是常有的是，今天仍然有人过着这样的生活。 在仅有的一个房间哩，如果要摆下全家人的床铺，就不能在那里吃饭了。这一点，褥子很方便。早晨，不需要褥子的时候，可以收起来。在没有了褥子的房间放上桌子，当作饭厅吃早饭。来客人的话，就在那里喝茶；孩子放学回到家里，那房间就成了书房。而后，傍晚又成为饭厅。然后收起桌子，铺上褥子，又成为了全家人睡觉的地方。 如果是床铺的话，除了睡觉的房间，还需要吃饭的房间和书房等，但如果使用褥子，一个房间就可以有各种用途。 据说从前，在纽约等大都市的大学学习的学生也租得起很大的房间。但现在房租太贵，租不起了。只能住更便宜、更小的房间。因此，似乎开始使用睡觉时作床，白天折小能成为椅子的、方便的褥子。
新视野大学英语第一册Unit 1课文翻译 学习外语是我一生中最艰苦也是最有意义的经历之一。 虽然时常遭遇挫折，但却非常有价值。 我学外语的经历始于初中的第一堂英语课。 老师很慈祥耐心，时常表扬学生。 由于这种积极的教学方法，我踊跃回答各种问题，从不怕答错。 两年中，我的成绩一直名列前茅。 到了高中后，我渴望继续学习英语。然而，高中时的经历与以前大不相同。 以前，老师对所有的学生都很耐心，而新老师则总是惩罚答错的学生。 每当有谁回答错了，她就会用长教鞭指着我们，上下挥舞大喊：“错！错！错！” 没有多久，我便不再渴望回答问题了。 我不仅失去了回答问题的乐趣，而且根本就不想再用英语说半个字。 好在这种情况没持续多久。 到了大学，我了解到所有学生必须上英语课。 与高中老师不同，大学英语老师非常耐心和蔼，而且从来不带教鞭！ 不过情况却远不尽如人意。 由于班大，每堂课能轮到我回答的问题寥寥无几。 上了几周课后，我还发现许多同学的英语说得比我要好得多。 我开始产生一种畏惧感。 虽然原因与高中时不同，但我却又一次不敢开口了。 看来我的英语水平要永远停步不前了。 直到几年后我有机会参加远程英语课程，情况才有所改善。 这种课程的媒介是一台电脑、一条电话线和一个调制解调器。 我很快配齐了必要的设备并跟一个朋友学会了电脑操作技术，于是我每周用5到7天在网上的虚拟课堂里学习英语。 网上学习并不比普通的课堂学习容易。 它需要花许多的时间，需要学习者专心自律，以跟上课程进度。 我尽力达到课程的最低要求，并按时完成作业。 我随时随地都在学习。 不管去哪里，我都随身携带一本袖珍字典和笔记本，笔记本上记着我遇到的生词。 我学习中出过许多错，有时是令人尴尬的错误。 有时我会因挫折而哭泣，有时甚至想放弃。 但我从未因别的同学英语说得比我快而感到畏惧，因为在电脑屏幕上作出回答之前，我可以根据自己的需要花时间去琢磨自己的想法。 突然有一天我发现自己什么都懂了，更重要的是，我说起英语来灵活自如。 尽管我还是常常出错，还有很多东西要学，但我已尝到了刻苦学习的甜头。 学习外语对我来说是非常艰辛的经历，但它又无比珍贵。 它不仅使我懂得了艰苦努力的意义，而且让我了解了不同的文化，让我以一种全新的思维去看待事物。 学习一门外语最令人兴奋的收获是我能与更多的人交流。 与人交谈是我最喜欢的一项活动，新的语言使我能与陌生人交往，参与他们的谈话，并建立新的难以忘怀的友谊。 由于我已能说英语，别人讲英语时我不再茫然不解了。 我能够参与其中，并结交朋友。
习惯与礼仪 我是个漫画家，对旁人细微的动作、不起眼的举止等抱有好奇。所以，我在国外只要做错一点什么，立刻会比旁人更为敏锐地感觉到那个国家的人们对此作出的反应。 譬如我多次看到过，欧美人和中国人见到我们日本人吸溜吸溜地出声喝汤而面露厌恶之色。过去，日本人坐在塌塌米上，在一张低矮的食案上用餐，餐具离嘴较远。所以，养成了把碗端至嘴边吸食的习惯。喝羹匙里的东西也象吸似的，声声作响。这并非哪一方文化高或低，只是各国的习惯、礼仪不同而已。 日本人坐在椅子上围桌用餐是1960年之后的事情。当时，还没有礼仪规矩，甚至有人盘着腿吃饭。外国人看见此景大概会一脸厌恶吧。 韩国女性就座时，单腿翘起。我认为这种姿势很美，但习惯于双膝跪坐的日本女性大概不以为然，而韩国女性恐怕也不认为跪坐为好。 日本等多数亚洲国家，常有人习惯在路上蹲着。欧美人会联想起狗排便的姿势而一脸厌恶。 日本人常常把手放在小孩的头上说“好可爱啊！”，而大部分外国人会不愿意。 如果向回教国家的人们劝食猪肉和酒，或用左手握手、递东西，会不受欢迎的。当然，饭菜也用右手抓着吃。只有从公用大盘往自己的小盘里分食用的公勺是用左手拿。一旦搞错，用黏糊糊的右手去拿，
会遭人厌恶。 在欧美，对不受欢迎的客人不说“请脱下外套”，所以电视剧中的侦探哥隆波总是穿着外套。访问日本家庭时，要在门厅外脱掉外套后进屋。穿到屋里会不受欢迎的。 这些习惯只要了解就不会出问题，如果因为不知道而遭厌恶、憎恨，实在心里难受。 过去，我曾用色彩图画和简短的文字画了一本《关键时刻的礼仪》（新潮文库）。如今越发希望用各国语言翻译这本书。以便能对在日本的外国人有所帮助。同时希望有朝一日以漫画的形式画一本“世界各国的习惯与礼仪”。 练习答案 ５、 （１）止める並んでいる見ているなる着色した （２）拾った入っていた行ったしまった始まっていた
新视野大学英语第三版第二册读写课文翻译 Unit 1 Text A 一堂难忘的英语课 1 如果我是唯一一个还在纠正小孩英语的家长，那么我儿子也许是对的。对他而言，我是一个乏味的怪物：一个他不得不听其教诲的父亲，一个还沉湎于语法规则的人，对此我儿子似乎颇为反感。 2 我觉得我是在最近偶遇我以前的一位学生时，才开始对这个问题认真起来的。这个学生刚从欧洲旅游回来。我满怀着诚挚期待问她：“欧洲之行如何？” 3 她点了三四下头，绞尽脑汁，苦苦寻找恰当的词语，然后惊呼：“真是，哇！” 4 没了。所有希腊文明和罗马建筑的辉煌居然囊括于一个浓缩的、不完整的语句之中！我的学生以“哇！”来表示她的惊叹，我只能以摇头表达比之更强烈的忧虑。 5 关于正确使用英语能力下降的问题，有许多不同的故事。学生的确本应该能够区分诸如their/there/they're之间的不同，或区别complimentary 跟complementary之间显而易见的差异。由于这些知识缺陷，他们承受着大部分不该承受的批评和指责，因为舆论认为他们应该学得更好。 6 学生并不笨，他们只是被周围所看到和听到的语言误导了。举例来说，杂货店的指示牌会把他们引向stationary（静止处），虽然便笺本、相册、和笔记本等真正的stationery（文具用品）并没有被钉在那儿。朋友和亲人常宣称They've just ate。实际上，他们应该说They've just eaten。因此，批评学生不合乎情理。 7 对这种缺乏语言功底而引起的负面指责应归咎于我们的学校。学校应对英语熟练程度制定出更高的标准。可相反，学校只教零星的语法，高级词汇更是少之又少。还有就是，学校的年轻教师显然缺乏这些重要的语言结构方面的知识，因为他们过去也没接触过。学校有责任教会年轻人进行有效的语言沟通，可他们并没把语言的基本框架——准确的语法和恰当的词汇——充分地传授给学生。
新视野大学英语1课文翻译 1下午好！作为校长，我非常自豪地欢迎你们来到这所大学。你们所取得的成就是你们自己多年努力的结果，也是你们的父母和老师们多年努力的结果。在这所大学里，我们承诺将使你们学有所成。 2在欢迎你们到来的这一刻，我想起自己高中毕业时的情景，还有妈妈为我和爸爸拍的合影。妈妈吩咐我们：“姿势自然点。”“等一等,”爸爸说，“把我递给他闹钟的情景拍下来。”在大学期间，那个闹钟每天早晨叫醒我。至今它还放在我办公室的桌子上。 3让我来告诉你们一些你们未必预料得到的事情。你们将会怀念以前的生活习惯，怀念父母曾经提醒你们要刻苦学习、取得佳绩。你们可能因为高中生活终于结束而喜极而泣，你们的父母也可能因为终于不用再给你们洗衣服而喜极而泣！但是要记住：未来是建立在过去扎实的基础上的。 4对你们而言，接下来的四年将会是无与伦比的一段时光。在这里，你们拥有丰富的资源：有来自全国各地的有趣的学生，有学识渊博又充满爱心的老师，有综合性图书馆，有完备的运动设施，还有针对不同兴趣的学生社团——从文科社团到理科社团、到社区服务等等。你们将自由地探索、学习新科目。你们要学着习惯点灯熬油，学着结交充满魅力的人，学着去追求新的爱好。我想鼓励你们充分利用这一特殊的经历，并用你们的干劲和热情去收获这一机会所带来的丰硕成果。 5有这么多课程可供选择，你可能会不知所措。你不可能选修所有的课程，但是要尽可能体验更多的课程！大学里有很多事情可做可学，每件事情都会为你提供不同视角来审视世界。如果我只能给你们一条选课建议的话，那就是：挑战自己！不要认为你早就了解自己对什么样的领域最感兴趣。选择一些你从未接触过的领域的课程。这样，你不仅会变得更加博学，而且更有可能发现一个你未曾想到的、能成就你未来的爱好。一个绝佳的例子就是时装设计师王薇薇，她最初学的是艺术史。随着时间的推移，王薇薇把艺术史研究和对时装的热爱结合起来，并将其转化为对设计的热情，从而使她成为全球闻名的设计师。 6在大学里，一下子拥有这么多新鲜体验可能不会总是令人愉快的。在你的宿舍楼里，住在你隔壁寝室的同学可能会反复播放同一首歌，令你头痛欲裂！你可能喜欢早起，而你的室友却是个夜猫子！尽管如此，你和你的室友仍然可能成
新视野大学英语2课文翻译（Unit1-Unit7） Unit 1 Section A 时间观念强的美国人 Para. 1 美国人认为没有人能停止不前。如果你不求进取，你就会落伍。这种态度造就了一个投身于研究、实验和探索的民族。时间是美国人注意节约的两个要素之一，另一个是劳力。 Para. 2 人们一直说：“只有时间才能支配我们。”人们似乎是把时间当作一个差不多是实实在在的东西来对待的。我们安排时间、节约时间、浪费时间、挤抢时间、消磨时间、缩减时间、对时间的利用作出解释；我们还要因付出时间而收取费用。时间是一种宝贵的资源，许多人都深感人生的短暂。时光一去不复返。我们应当让每一分钟都过得有意义。 Para. 3 外国人对美国的第一印象很可能是：每个人都匆匆忙忙——常常处于压力之下。城里人看上去总是在匆匆地赶往他们要去的地方，在商店里他们焦躁不安地指望店员能马上来为他们服务，或者为了赶快买完东西，用肘来推搡他人。白天吃饭时人们也都匆匆忙忙，这部分地反映出这个国家的生活节奏。工作时间被认为是宝贵的。Para. 3b 在公共用餐场所，人们都等着别人吃完后用餐，以便按时赶回去工作。你还会发现司机开车很鲁莽，人们推搡着在你身边过去。你会怀念微笑、简短的交谈以及与陌生人的随意闲聊。不要觉得这是针对你个人的，这是因为人们非常珍惜时间，而且也不喜欢他人“浪费”时间到不恰当的地步。 Para. 4 许多刚到美国的人会怀念诸如商务拜访等场合开始时的寒暄。他们也会怀念那种一边喝茶或咖啡一边进行的礼节性交流，这也许是他们自己国家的一种习俗。他们也许还会怀念在饭店或咖啡馆里谈生意时的那种轻松悠闲的交谈。一般说来，美国人是不会在如此轻松的环境里通过长时间的闲聊来评价他们的客人的，更不用说会在增进相互间信任的过程中带他们出去吃饭，或带他们去打高尔夫球。既然我们通常是通过工作而不是社交来评估和了解他人，我们就开门见山地谈正事。因此，时间老是在我们心中的耳朵里滴滴答答地响着。 Para. 5 因此，我们千方百计地节约时间。我们发明了一系列节省劳力的装置；我们通过发传真、打电话或发电子邮件与他人迅速地进行交流，而不是通过直接接触。虽然面对面接触令人愉快，但却要花更多的时间, 尤其是在马路上交通拥挤的时候。因此，我们把大多数个人拜访安排在下班以后的时间里或周末的社交聚会上。 Para. 6 就我们而言，电子交流的缺乏人情味与我们手头上事情的重要性之间很少有或完全没有关系。在有些国家, 如果没有目光接触，就做不成大生意，这需要面对面的交谈。在美国，最后协议通常也需要本人签字。然而现在人们越来越多地在电视屏幕上见面，开远程会议不仅能解决本国的问题，而且还能通过卫星解决国际问题。
新视野大学英语3第三版课文翻译 Unit 1 The Way to Success 课文A Never, ever give up! 永不言弃！ As a young boy, Britain's great Prime Minister, Sir Winston Churchill, attended a public school called Harrow. He was not a good student, and had he not been from a famous family, he probably would have been removed from the school for deviating from the rules. Thankfully, he did finish at Harrow and his errors there did not preclude him from going on to the university. He eventually had a premier army career whereby he was later elected prime minister. He achieved fame for his wit, wisdom, civic duty, and abundant courage in his refusal to surrender during the miserable dark days of World War II. His amazing determination helped motivate his entire nation and was an inspiration worldwide. Toward the end of his period as prime minister, he was invited to address the patriotic young boys at his old school, Harrow. The headmaster said, "Young gentlemen, the greatest speaker of our time, will be here in a few days to address you, and you should obey whatever sound advice he may give you." The great day arrived. Sir Winston stood up, all five feet, five inches and 107 kilos of him, and gave this short, clear-cut speech: "Young men, never give up. Never give up! Never give up! Never, never, never, never!" 英国的伟大首相温斯顿·丘吉尔爵士，小时候在哈罗公学上学。当时他可不是个好学生，要不是出身名门，他可能早就因为违反纪律被开除了。谢天谢地，他总算从哈罗毕业了，在那里犯下的错误并没影响到他上大学。后来，他凭着军旅生涯中的杰出表现当选为英国首相。他的才思、智慧、公民责任感以及在二战痛苦而黑暗的时期拒绝投降的无畏勇气，为他赢得了美名。他非凡的决心，不仅激励了整个民族，还鼓舞了全世界。 在他首相任期即将结束时，他应邀前往母校哈罗公学，为满怀报国之志的同学们作演讲。校长说：“年轻的先生们，当代最伟大的演说家过几天就会来为你们演讲，他提出的任何中肯的建议，你们都要听从。”那个激动人心的日子终于到了。温斯顿爵士站了起来——他只有5 英尺5 英寸高，体重却有107 公斤。他作了言简意赅的讲话：“年轻人，要永不放弃。永不放弃！永不放弃！永不，永不，永不，永不！” Personal history, educational opportunity, individual dilemmas - none of these can inhibit a strong spirit committed to success. No task is too hard. No amount of preparation is too long or too difficult. Take the example of two of the most scholarly scientists of our age, Albert Einstein and Thomas Edison. Both faced immense obstacles and extreme criticism. Both were called "slow to learn" and written off as idiots by their teachers. Thomas Edison ran away from school because his teacher whipped him repeatedly for asking too many questions. Einstein didn't speak fluently until he was almost nine years old and was such a poor student that some thought he was unable to learn. Yet both boys' parents believed in them. They worked intensely each day with their sons, and the boys learned to never bypass the long hours of hard work that they needed to succeed. In the end, both Einstein and Edison overcame their childhood persecution and went on to achieve magnificent discoveries that benefit the entire world today. Consider also the heroic example of Abraham Lincoln, who faced substantial hardships, failures and repeated misfortunes in his lifetime. His background was certainly not glamorous. He was raised in a very poor family with only one year of formal education. He failed in business twice, suffered a nervous breakdown when his first love died suddenly and lost eight political
Text A 课文 A The humanities: Out of date? 人文学科：过时了吗？ When the going gets tough, the tough takeaccounting. When the job market worsens, manystudents calculate they can't major in English orhistory. They have to study something that booststheir prospects of landing a job. 当形势变得困难时，强者会去选学会计。当就业市场恶化时，许多学生估算着他们不能再主修英语或历史。他们得学一些能改善他们就业前景的东西。 The data show that as students have increasingly shouldered the ever-rising c ost of tuition,they have defected from the study of the humanities and toward applied science and "hard"skills that they bet will lead to employment. In oth er words, a college education is more andmore seen as a means for economic betterment rather than a means for human betterment.This is a trend that i s likely to persist and even accelerate. 数据显示，随着学生肩负的学费不断增加，他们已从学习人文学科转向他们相信有益于将来就业的应用科学和“硬”技能。换言之，大学教育越来越被看成是改善经济而不是提升人类自身的手段。这种趋势可能会持续，甚至有加快之势。 Over the next few years, as labor markets struggle, the humanities will proba bly continue theirlong slide in succession. There already has been a nearly 50 percent decline in the portion of liberal arts majors over the past generatio n, and it is logical to think that the trend is boundto continue or even accel erate. Once the dominant pillars of university life, the humanities nowplay li ttle roles when students take their college tours. These days, labs are more vi vid and compelling than libraries. 在未来几年内，由于劳动力市场的不景气，人文学科可能会继续其长期低迷的态势。在上一代大学生中，主修文科的学生数跌幅已近50%。这种趋势会持续、甚至加速的想法是合情合理的。人文学科曾是大学生活的重要支柱，而今在学生们参观校园的时候，却只是一个小点缀。现在，实验室要比图书馆更栩栩如生、受人青睐。 Here, please allow me to stand up for and promote the true value that the h umanities add topeople's lives. 在这儿，请允许我为人文学科给人们的生活所增添的真实价值进行支持和宣传。
第10课 日本的季节 日本的一年有春、夏、秋、冬四个季节。 3月、4月和5月这三个月是春季。春季是个暖和的好季节。桃花、樱花等花儿开得很美。人们在4月去赏花。 6月到8月是夏季。夏季非常闷热。人们去北海道旅游。7月和8月是暑假，年轻人去海边或山上。也有很多人去攀登富士山。富士山是日本最高的山。 9月、10月和11月这3个月是秋季。秋季很凉爽，晴朗的日子较多。苹果、桔子等许多水果在这个季节成熟。 12月到2月是冬季。日本的南部冬天不太冷。北部非常冷，下很多雪。去年冬天东京也很冷。今年大概不会那么冷吧。如果冷的话，人们就使用暖气炉。 第12课 乡下 我爷爷住哎乡下。今天，我要去爷爷家。早上天很阴，但中午天空开始变亮，天转好了。我急急忙忙吃完午饭，坐上了电车。 现在，电车正行驶在原野上。窗外，水田、旱地连成一片。汽车在公路上奔驰。 这时，电车正行驶在大桥上。下面河水在流动。河水很清澈，可以清澈地看见河底。可以看见鱼在游动。远处，一个小孩在挥手。他身旁，牛、马在吃草。 到了爷爷居住的村子。爷爷和奶奶来到门口等着我。爷爷的房子是旧房子，但是很大。登上二楼，大海就在眼前。海岸上，很多人正在全力拉缆绳。渐渐地可以看见网了。网里有很多鱼。和城市不同，乡下的大自然真是很美。 第13课 暑假 大概没有什么比暑假更令学生感到高兴的了。大学在7月初，其他学校在二十四日左右进入暑假。暑假大约1个半月。 很多人利用这个假期去海边、山上，或者去旅行。学生中，也有人去打工。学生由于路费等只要半价，所以在学期间去各地旅行。因此，临近暑假时，去北海道的列车上就挤满了这样的人。从炎热的地方逃避到凉爽的地方去，这是很自然的事。一般在1月、最迟在2月底之前就要预定旅馆。不然的话可能会没有地方住。 暑假里，山上、海边、湖里、河里会出现死人的事，这种事故都是由于不注意引起的。大概只能每个人自己多加注意了。 在东京附近，镰仓等地的海面不起浪，因此挤满了游泳的人。也有人家只在夏季把海边的房子租下来。 暑假里，学校的老师给学生布置作业，但是有的学生叫哥哥或姐姐帮忙。 第14课 各式各样的学生 我就读的大学都有各种各样的学生入学。学生有的是中国人，有的是美国人，有的是英国人。既有年轻的，也有不年轻的。有胖的学生，也有瘦的学生。学生大多边工作边学习。因此，大家看上去都很忙。经常有人边听课边打盹。 我为了学习日本先进的科学技术和日本文化来到日本。预定在这所大学学习3年。既然特意来了日本，所以每天都很努力学习。即便如此，考试之前还是很紧张。其他学生也是这
教育界的科技革命 如果让生活在年的人来到我们这个时代，他会辨认出我们当前课堂里发生的许多事情——那盛行的讲座、对操练的强调、从基础读本到每周的拼写测试在内的教学材料和教学活动。可能除了教堂以外，很少有机构像主管下一代正规教育的学校那样缺乏变化了。 让我们把上述一贯性与校园外孩子们的经历作一番比较吧。在现代社会，孩子们有机会接触广泛的媒体，而在早些年代这些媒体简直就是奇迹。来自过去的参观者一眼就能辨认出现在的课堂，但很难适应现今一个岁孩子的校外世界。 学校——如果不是一般意义上的教育界——天生是保守的机构。我会在很大程度上为这种保守的趋势辩护。但变化在我们的世界中是如此迅速而明确，学校不可能维持现状或仅仅做一些表面的改善而生存下去。的确，如果学校不迅速、彻底地变革，就有可能被其他较灵活的机构取代。 计算机的变革力 当今时代最重要的科技事件要数计算机的崛起。计算机已渗透到我们生活的诸多方面，从交通、电讯到娱乐等等。许多学校当然不能漠视这种趋势，于是也配备了计算机和网络。在某种程度上，这些科技辅助设施已被吸纳到校园生活中，尽管他们往往只是用一种更方便、更有效的模式教授旧课程。 然而，未来将以计算机为基础组织教学。计算机将在一定程度上允许针对个人的授课，这种授课形式以往只向有钱人提供。所有的学生都会得到符合自身需要的、适合自己学习方法和进度的课程设置，以及对先前所学材料、课程的成绩记录。 毫不夸张地说，计算机科技可将世界上所有的信息置于人们的指尖。这既是幸事又是灾难。我们再也无须花费很长时间查找某个出处或某个人——现在，信息的传递是瞬时的。不久，我们甚至无须键入指令，只需大声提出问题，计算机就会打印或说出答案，这样，人们就可实现即时的"文化脱盲"。 美中不足的是，因特网没有质量控制手段；"任何人都可以拨弄"。信息和虚假信息往往混杂在一起，现在还没有将网上十分普遍的被歪曲的事实和一派胡言与真实含义区分开来的可靠手段。要识别出真的、美的、好的信息，并挑出其中那些值得知晓的, 这对人们构成巨大的挑战。 对此也许有人会说，这个世界一直充斥着错误的信息。的确如此，但以前教育当局至少能选择他们中意的课本。而今天的形势则是每个人都拥有瞬时可得的数以百万计的信息源，这种情况是史无前例的。 教育的客户化 与以往的趋势不同，从授权机构获取证书可能会变得不再重要。每个人都能在模拟的环境中自学并展示个人才能。如果一个人能像早些时候那样"读法律"，然后通过计算机模拟的实践考试展现自己的全部法律技能，为什么还要花万美元去上法学院呢？用类似的方法学开飞机或学做外科手术不同样可行吗？ 在过去，大部分教育基本是职业性的：目的是确保个人在其年富力强的整个成人阶段能可靠地从事某项工作。现在，这种设想有了缺陷。很少有人会一生只从事一种职业；许多人都会频繁地从一个职位、公司或经济部门跳到另一个。 在经济中，这些新的、迅速变换的角色的激增使教育变得大为复杂。大部分老成持重的教师和家长对帮助青年一代应对这个会经常变换工作的世界缺乏经验。由于没有先例，青少年们只有自己为快速变化的"事业之路"和生活状况作准备。