P R O P O R T I O N A L I T Y G R A P H S, U N I T S A N A L Y S I S, A N D D O M A I N C O N S T R A I N T S:
I M P R O V I N G T H E P O W E R A N D E F F I C I E N C Y O F T H E S C I E N T I F I C D I S C O V E R Y PROCESS*
Brian Falkenhainer
Department of Computer Science
University of Illinois
1304 W. Springfield Ave, Urbana, IL 61801
A B S T R A C T
An i m p o r t a n t subproblem of scientific discovery is quantitative discovery, finding formulas t h a t relate some set (or subset) of a collection of numerical parameters. Current work in quantitative discovery suffers from a lack of efficiency and generality. This paper discusses methods that are efficient and yet general for discovering equations which try to avoid exponential search. Importantly, these methods can derive equations t h a t cover subsets of the data and derive explicit descriptions of when the equations are applicable. These methods are fully implemented in a system named ABACUS which is described and some of its results are presented.
1. I N T R O D U C T I O N
A goal of quantitative scientific discovery is to create systems which will be able to propose, more or less on their
o w n, various types of empirical laws which hold for some domain. Such systems should be both general and efficient, should be able to handle irrelevant variables, and should describe, where possible, the limits of applicability for their conclusions. While some progress had been made, most notably the BACON series [Langley 83], existing systems have severe drawbacks. These systems tend to make various assumptions which only enable them to discover laws that hold for roughly all of the given data. They assume t h a t all variables are relevant and also require the user to specify which variables should be treated as independent and which as dependent. This is particularly limiting since in interesting cases we seldom know much about the variables involved.
This paper presents several methods to help overcome these limitations. First, t w o methods of constraining the search, proportionality graphs and units analysis, are discussed. Second, a method for deriving several empirical laws t h a t cover different portions of a dataset is introduced
t h a t produces a description of when the empirical law is applicable. Finally, some results generated by the program ABACUS are shown.
S. C O N S T R A I N I N G 8E A R C H
T h e search for empirical laws can be roughly characterized as searching through the space of equations which can be generated from the variables given in the dataset, stopping whenever a law that accounts for a significant amount of the dataset is found. T w o problems confound this search. First, the presence of irrelevant variables in the data can tend to confuse the search process. Second, the process of combining existing terms to form new terms is inherently combinatorial. A t t e m p t s have been made to constrain the process using "expectation-driven heuristics" which may use desired formula templates, knowledge about the units of measurement being used, or the domain of the problem in general. An approach which makes no assumptions would be of more value.
2.1. P r o p o r t i o n a l i t y Graphs
T w o variables, x and y, may be said to be qualitatively proportional if, when other variables are held constant, x rises monotonically as y rises. They are said to be inversely qualitatively proportional if x decreases monotonically as y rises. There are four assertions possible as the result of a qualitative proportionality measurement:
prop+(x,y) - x and y are qualitatively proportional
p r o p-(x,y) - x and y arc inversely qualitatively proportional noprop(x.y) - x and y are unrelated
prop? (x. y) - insufficient data to determine if x and y are related
Once all existing variables have been so related, an undirected graph, called the proportionality graph, is constructed where nodes represent variables and an edge indicates the existence of a proportionality relation ( + or -) between two variables. It is postulated t h a t, for formulas consisting solely of multiplication and division, the relevant variables in the formula will appear in a cycle in the proportionality graph. This is due to the fact that for formulas of the form (xyz/abc = constant), each variable's value will appear to be dependent upon the other 5. For other formula types, which may or may not adhere to the cycle rule, the most promising area for search would be the largest connected component of the graph.
For example, given data of 6 variables (P, V, N, T, M, Y) where the ideal gas law is the intended relation (P V/N T = 8.32), the proportionality graph might look like:
* This work was supported in part by the National Science Founda-tion under grant NSF DCR 84-06801 and by the Office of Naval Research under grant no. N00014-82-K-0186.
B. Falkenhalner 553
A simple depth-first search on the relations existing among the set of variables in the cycle {P, V, N, T} produces the desired relation*. No further examination of M or Y are needed for this example. If the quick depth-first search failed, then variables corresponding to edges prop+N.M) and prop±(V,Y) would have been created and search would have continued. This procedure removes irrelevant variables whenever possible and directs the search towards the most promising relations in the data.
2.2. U n i t s Analysis
Once it has been decided to create new variables for the relation p r o p-(x.y), ABACUS tries to form both sum relations (e.g. x+y) and a product relation (xXy) in an effort to create a variable w i t h a constant value. W i t h o u t further constraint, creating two or more new variables for each proportionality assertion might cause the number of variables in the system to grow exponentially with the search depth. A simple physical constraint drastically reduces the number of combinations that must be considered. For two entities to be added or subtracted, they must be the same type of entity, that is, they must be in the same units. One may divide meters by seconds, but not subtract seconds from meters. Thus, when an attempt is made to form x+y, the units of x and y are examined for compatibility. If x and y have the same units, x+y and x2+y2 are created (a need for higher powers has not presented itself as yet). If the units of x are a power of y's units, then y is raised to the appropriate power (e.g. x + y2). Notice we are only testing identity of units and no semantic interpretation is done to guide the search.
3. G E N E R A L I Z I N G T H E D I S C O V E R I E S
One of the goals of scientific discovery so far overlooked by past systems is characterizing the domain in which the laws generated are relevant. Only one law could be found for each data set - unlike real life, these programs have not been tested on data sets which include several independent or semi-independent relationships. The ability to classify data according to what mathematical formulae best describes it will be very important as more realistic problems are undertaken.
In the ABACUS system, when the best relation found describes only a portion of the given data, the events covered by this relation are removed from the data and placed in a unique set. The search process begins again to find relationships t h a t describe the remaining events. Entirely new relations may be found this time since the removed variables may have been dominating the relation finding process d u r i n g the first pass. This process repeats until all of the data has been functionally described or a set of events remains for which no rule can be found. Once all the events are so analyzed, the system attempts to discover what factors present in the data caused set i to be different from set j by using an algorithm known as Aq [Michalski 83]. Given event sets such as those created above, Aq will generate a description for each set which covers the examples of that set
* The correct term here is biconnected component [Aho 76] which refers to all nodes forming a cycle - thus {P V N T} rather than merely {V N T}. and none of the examples in any other class. This is called a discriminant description in that it may be used to determine to which class a given event belongs. (Aq is described in greater detail in [Michalski 83])
An example will make the role of Aq in the discovery
process clearer.
Data corresponding to the above graph was presented to ABACUS. After failing to locate a relation which held for all of the given data, search terminated w i t h y = x2 being the equation which held for the largest percentage of the data . A l l data covered by this formula was removed from the dataset and search began again. This time the equation describing the line was discovered and found to hold universally for all of the current data. W i t h relations found for the entire dataset, Aq was then called to discriminate
between the two sets of events. The results are shown below:
4. I M P L E M E N T A T I O N
The concepts presented above are implemented in the ABACUS system, which is written in FranzLisp. ABACUS operates in two stages. First, the given data is examined for empirical rules which hold over different subsets of the events. Once all of the data has been separated into disjoint classes corresponding to each empirical rule, the second phase begins. This stage applies the Aq algorithm to the classes to create discriminate descriptions of the class sets. The class description and empirical rule which describes t h a t class are then combined to form the if-then rules illustrated above.
Since we have already described the role of Aq in some detail, we now describe how the first stage works. First the system searches through the data and produces all possible proportionality assertions. A simple graph traversal routine is then called to return an ordered list of the cycles and connected components of the proportionality graph. A depth-first search is performed on each set in t u r n, where the
It should be pointed out that the problem of conflicting propor-tionalities has not been solved here. For the curve, prop*(x,y) holds, while prop-(x,y) holds for the line.. In this example, there were 16 points given for the curve and 7 given for the line. ABACUS is able to solve prob-lems having conflicting proportionalities if one dominates the other, as in this example.
554 B. Falkenhainer
depth of the search for a set is given by the cardinality of the set. Variables, once created, are added to the events list and are never retracted. If a relationship t h a t describes all the data is not f o u n d , search proceeds in a best-first manner until such a universal relation is found or a cutoff parameter is reached. If no universal relation is found, the events covered by the relation which summarizes the most events are removed from the data, placed in a unique class set, and all generated variables are t h r o w n out. The process repeats on the remaining data. 4.1. A F e w R e s u l t s
One law which has been discovered by the current ABACUS system is Newton's conservation of linear kinetic energy in a perfectly elastic collision. This law states t h a t the total translational kinetic energy of t w o bodies remains
the same before and after an elastic collision:
In this example, ABACUS was presented w i t h data consisting of the eight variables shown above and a nominal collision-type variable which took on values of ''elastic" and "inelastic". As can be seen below, the system was able to discover the formula for conservation of kinetic energy. Since this relation only held for about half of the data, further search was attempted on the remaining half, but to no avail. A domain constraint was therefore provided to discriminate the t w
o halves of the data.
class-b: If [collision-type = inelastic]
then No formula holds
Another interesting example which helps to show the usefulness of the domain constraint concept involves Coulomb's law of electrical force. The electrical force
between two charged objects is given by
where c is the electrical p e r m i t t i v i t y of the surrounding medium. ABACUS was presented w i t h data consisting of the force, the t w o charges, the distance between the objects, and the name of the surrounding medium. The results are given here, where the constant term in each of the Coulomb relations is equal to for t h a
t substance.
5. C O N C L U S I O N
The ABACUS system is a very useful exercise in simulating a part of the scientific discovery process. It has been designed to be efficient and general and has been shown to be applicable to a variety of chemical and physical laws. This combination of t w o completely separate algorithms (the relation finding process and Aq) shows how different approaches to learning can be combined to enable tasks neither could do alone. In experimentation, the pertinent variables are often mixed in w i t h many completely irrelevant variables. W i t h this design, a wide variety of variables may be introduced and decisions of pertinence may be left to the program. W i t h Aq acting as a post-processor, the data can then be examined to answer questions concerning the applicable domain of the results and postulate the reasons for discovered differences in the data.
There are still a number of unresolved problems in quantitative discovery. ABACUS is unable to find laws involving logarithmic or trigonometric relations unless these functions are directly introduced into the data by the user. Also unanswered is the problem of discovering polynomials w i t h coefficients. While a great many laws of physics and chemistry do not contain coefficients in their equations, many important laws do, particularly Newton's laws of motion.
A C K N O W L E D G E M E N T S
I would like to thank my advisor, Ryszard Michalski, for his support and advice on this project. I also owe a great deal of gratitude to K en Forbus, Tony Nowicki, Jeff Becker, and Sue Crane for their time, suggestions, and overall help.
R E F E R E N C E S
[1] Aho, A., J.E.Hopcroft, J.D.UIlman, The Design and Analysis of Computer Algorithms, Addison-Wesley, 1976. [2] de K leer,J., "Qualitative and Quantitative K nowledge in Classical Mechanics," MS Thesis, TR-352, Massachusetts Institute of Technology, Cambridge, M A , 1975.
[3] Falkenhainer,B., "A B A C U S : A d d i n g Domain Constraints to Quantitative Scientific Discovery," UIUCDCS-F-84-927, ISO 84-7, University of Illinois, November 1984.
[4] Forbus, K .D., "Qualitative Process Theory," PhD Thesis, TR-798, Massachusetts Institute of Technology, Cambridge, M A , 1984.
[5] Langley,P., G.L.Bradshaw, H.A.Simon, "Rediscovering Chemistry w i t h the Bacon System." In R.S.Michalski, J.G.Carbonell, and T.M.M i t c h e l l (Eds.), Machine Learning -An Artificial Intelligence Approach. Tioga Pub., 1983.
[6] LangIey,P., G.L.Bradshaw, H.A.Simon, J.Zytkow, "Mechanisms for Qualitative and Quantitative Discovery," Proceedings of the International Machine Learning Workshop, Monticello, Illinois, 1983.
[7] Michalski,R.S., "A Comparative Review of Selected Methods for Learning from Examples." In R.S.Michalski, J.G.Carbonell, and T.M.M i t c h e l l (Eds.), Machine Learning -An Artificial Intelligence Approach. Tioga Pub., 1983.
名词解释 1.文艺学：研究文学及其规律的学科统称文艺学，它是一门以文学为研究对象，以揭示文学基本规律，介绍相关知识为目的的学科，属于人文学科的畴。 2.文学理论：文艺学的分支之一，以文学的基本原理、概念、畴以及相关的研究法为研究对象，并将分析研究文学的普遍规律作为其根本任务。 3.文本：文本是指由作者写成，而有待于阅读的单个文学作品本身。艾布拉姆斯提出文学作为一种活动是由作品、作家、世界、读者等四个要素组成的。 4.文学价值：文学价值是文学作品满足人和社会需要的属性，是由作家和读者共同创造的，主客观统一的产物，主要包括认识价值、伦理价值、审美价值等。 5.文学价值的“真”：文学价值的真，是指文学要通过合乎艺术规律的式，将社会的真实状况、人生的真正面目、作家的真诚体验等表现出来。 6.文学价值的“美”：文学价值的美，是指文学在真和善相统一的基础上，满足人们对美的追求和需要，给人精神上的愉悦。 7.文学的功能定义：文学功能是文学价值属性的实际反映和体现。文学功能存在的在依据是文学的价值。文学的功能不是孤立地存在的，它存在于功能的系统之中。 8.文学创作（创作过程）：是指作家从产生创作动机和创作冲动到完
成艺术构思和艺术传达的过程。（这也是一部文学作品从在心理体验到外在形式的形成过程。） 语言呈现：是作家将构思成熟的艺术形象用语言表达出来的过程。（eg：眼中之竹到胸中之竹到手中之竹。眼中之竹山是指客观存在的竹子，而胸中之竹则是艺术构思过程中的艺术形象，至于手中之竹，已经是经过艺术传达之后的竹子，是画在纸上的画图，是最终完成的作品。） 国维的隔与不隔：“隔”意思则是指艺术技巧使用的很拙劣，以致不能有效地调动读者情感，并使之升华为艺术情感。“不隔”第一层含义是说文字运用得恰到好处，是读者能够直接体会到诗词中蕴含的涵而感不到丝毫的文字障碍。第二层含义是指用直书其事的式作诗，不堆砌典故，使人不劳猜想就直接感知诗中蕴含的情感。（池塘生春草不隔，“家池上，江淹浦畔”，隔） 文学创作过程（创作动因、艺术构思、语言呈现）、文学创作心理机制（艺术直觉、艺术情感、艺术想象、艺术理解）、文学创作的主体条件和追求（作家与生活体验、思想道德修养与文化艺术素养、创作个性与独创性、创作自由和社会责任） 创作动因：创作动因是指作家生活体验积累到一定程度时产生的创作驱力，包括创作动机和创作冲动。这是文学创作的开端。 9.创作动机：创作动机是指作家从事具体创作活动的目的。作家有某种思想感情需要传达，或要赞美、批评某种现存事物，或要互换某种新的社会变革等，心中有所积郁，不吐不快，于是产生了创作动机。
数学图形知识点 数学图形知识点 一、图形的变换 1、轴对称图形：把一个图形沿着某一条直线对折，两边能够完 全重合，这样的图形叫做轴对称图形，这条直线叫做对称轴。 2、成轴对称图形的特征和性质：①对称点到对称轴的距离相 等;②对称点的连线与对称轴垂直;③对称轴两边的图形大小形状完 全相同。 3、物体旋转时应抓住三点：①旋转中心;②旋转方向;③旋转角度。旋转只改变物体的位置，不改变物体的形状、大小。 二、因数与倍数 1、因数和倍数：如果整数a能被b整除，那么a就是b的倍数，b就是a的因数。 2、一个数的因数的求法：一个数的因数的个数是有限的，最小 的是1，最大的是它本身，方法是成对地按顺序找。 3、一个数的倍数的求法：一个数的倍数的个数是无限的，最小 的是它本身，没有最大的，方法时依次乘以自然数。 4、2、 5、3的倍数的特征：个位上是0、2、4、 6、8的数，都 是2的倍数。个位上是0或5的数，是5的倍数。一个数各位上的 数的和是3的倍数，这个数就是3的`倍数。 5、偶数与奇数：是2倍数的数叫做偶数(0也是偶数)，不是2 的倍数的数叫做奇数。
6、质数和和合数：一个数，如果只有1和它本身两个因数的数叫做质数(或素数)，最小的质数是2。一个数，如果除了1和它本身还有别的因数的数叫做合数，最小的合数是4。 三、长方体和正方体 1、长方体和正方体的特征：长方体有6个面，每个面都是长方形(特殊的有一组对面是正方形)，相对的面完全相同;有12条棱，相对的棱平行且相等;有8个顶点。正方形有6个面，每个面都是正方形，所有的面都完全相同;有12条棱，所有的棱都相等;有8个顶点。 2、长、宽、高：相交于一个顶点的三条棱的长度分别叫做长方体的长、宽、高。 3、长方体的棱长总和=(长+宽+高)×4正方体的棱长总和=棱长×12 4、表面积：长方体或正方体6个面的总面积叫做它的表面积。 5、长方体的表面积=(长×宽+长×高+宽×高)×2 S=(ab+ah+bh)×2 正方体的表面积=棱长×棱长×6 6、表面积单位：平方厘米、平方分米、平方米相邻单位的进率为100 7、体积：物体所占空间的大小叫做物体的体积。 8、长方体的体积=长×宽×高用字母表示：V=abh长=体积÷(宽×高)宽=体积÷(长×高) 高=体积÷(长×宽) 正方体的体积=棱长×棱长×棱长用字母表示：V=a×a×a 9、体积单位：立方厘米、立方分米和立方米相邻单位的进率为1000
一、填空题:（每空1 分，本大题共10 分） 1. （）语言学是在19世纪逐步发展和完善的，它是语言学 走上独立发展道路的标志。 2. 人的大脑分左右两半球，大脑的左半球控制（ 掌管不需要语言的感性直观思维。 3. 进入20世纪以后，语言研究的主流由历史比较语言学转为 （）。 4. 俄语属于印欧语系的（ 5. 一个音位包含的不同音素或者具体表现出来的音素叫做 （）。 6. 语言中最单纯、最常用、最原始和最能产的词是（ 7. 现代大多数国家的拼音文字的字母，大多直接来源于（）字 母。 8. 言外之意之所以能够被理解是因为（）起了补充说明的 作用。 9. 方言在社会完全分化的情况下，有可能发展成（? ）?； 在社会高度统一的情况下，会逐渐被共同语消磨直到同化。 10. 南京方言的“兰”、“南”不分，从音位变体的角度来说，[n ]和[l]是 属于（）变体。 二、单项选择题: 码填在题干上的括号内。（每小题1 分，本大题共15 分）
1. 在二十世纪，对哲学、人类学、心理学、社会学等学科产生重大影响 的语言学流派是（） A.历史比较语言学 B.心理语言学 C.结构主义语言学 D.社会语言学 2. “人有人言，兽有兽语”中的“言”属于（） A.语言 B.言语 C.言语行为 D.言语作品 3. “我爱家乡”中“爱”和“家乡”（） A.是聚合关系。 B.是组合关系。 C.既是聚合关系又是组合关系。 D. 4. 一种语言中数量最少的是 A.音素 B.音位 C.语素 D.音节 5. 英语的man—→men采用的语法手段是 A. 屈折变化 B.变换重音的位置 C. 变化中缀 D.异根 6. 在汉语普通话中没有意义区别功能的声学特征是（） A.音高 B.音强 C.音长 D.音质 7. [ε]的发音特征是 A.舌面前高不圆唇 B.舌面后高不圆唇 C.舌面前半高不圆唇 D.舌面前半低不圆唇 8. 构成“语言、身体”这两个词的语素的类型（） A.都是成词语素 B.都是不成词语素 C.“语”和“言”是成词语素，“身”和“体”是不成词语素 D.“语”和“言”是不成词语素，“身”和“体” 9. 广义地说，汉语动词词尾“着”、“了”、“过”属于语法范畴中的 （）
Chapter 6 Pragmatics 语用学 1.What is pragmatics? 什么是语用学？ Pragmatics can be defined as the study of how speakers of a language use sentences to effect successful communication. As the process of communication is essentially a process of conveying meaning in a certain context, pragmatics can also be regarded as a kind of meaning study. It places the study of meaning in the context in which language is used. 语用学研究的是说某种语言的人怎样用句子去实现成功的交际。 由于交际的过程从本质来说是在一定的语境中表达意义的过程，因而语用学的本质是一种意义研究。它是一种将语言置于使用的语境中去的意义研究。 2.Pragmatics and semantics 语用学和语义学 Pragmatics and semantics are both linguistic studies of meaning, but they are different. What essentially distinguishes semantics and pragmatics is whether in the study of meaning, the context of use is considered. If it is not considered, the study is restricted to the area of traditional semantics; if it is considered, the study is being carried out in the area of pragmatics. 语用学和语义学都是对意义的语言学研究，但两者是不同的。它们的本质区别在于研究意义时是否考虑了语言使用的语境。没有考虑到语境进行的研究就没有超出传统语义学的研究范围；相反，考虑到语境进行的研究就属于语用学的研究范围。 3.Context 语境 Context is essential to the pragmatic study of language. It is generally considered as constituted by the knowledge shared by the speaker and the hearer. 语境是语言的语用研究中不可缺少的概念。它一般被理解为说话者和听话者所共有的知识。The shared knowledge is of two types: the knowledge of the language they use, and the knowledge about the world, including the general knowledge about the world and the specific knowledge about the situation in which linguistic communication is taking place. 共有的知识包括他们所使用的语言方面的知识和双方对世界的认识，包括对世界的总的认识和对正在进行的语言交际所处的环境的具体认识。 4.Sentence meaning and utterance meaning 句子意义和话语意义The meaning of a sentence is abstract, and de-contextualized, while utterance meaning is concrete, and context-dependent. Utterance is based on sentence meaning; it is the realization of the abstract meaning of a sentence in a real situation of communication, or simply in a context. 句子的意义是抽象的，非语境化的，而话语的意义是具体的，受语境制约的。话语意义基于句子意义；它是一个句子的抽象意义在特定语境中的具体体现，或简而言之，在一个语境中的具体化。 5.Speech act theory 言语行为理论 Speech act theory is an important theory in the pragmatic study of language. It was originated with the British philosopher John Austin in the late 50’s of the 20th century. 言语行为理论是语言语用研究中的一个重要理论。它最初是由英国哲学家约翰.奥斯汀在20世纪50年代提出的。 According to speech act theory, we are performing actions when we are speaking.
1.再现论与表现论 再现论：文学是对现实的模仿，具有实践性。亚里士多德是再现论的奠基人。在中国，最早的文学作品是《二言弹歌》，“断竹续竹，飞土逐肉”再现当时人类的生存环境。 优点：①对文学艺术的源泉做出了唯物主义的解释。 ②十分重视文学艺术的认识价值。 ③引导作家深入生活。 缺点：①没有分清艺术与科学的界限，认为两者的本质都是给人以真知，只不过科学是以概念来把握生活，文学以形象来反映生活。 ②艺术创作的目的是为了逼近生活，以忠于现实，逼肖现实作为创作于批评 的最高标准，忽视作者对生活的情感体验。 表现论：文学是心灵的表现。如中国文论中意境的范畴，有我之境与无我之境。外国文论应追溯到柏拉图，他的“迷狂说”是表现论的理论基础，认为文学艺术陷入迷狂。真正对此理论作出贡献的是康德，认为知、情、意三者同样重要，作者主要表现情感。康德的表现论是是有审美标准的，理性的。此外还有弗洛伊德的“冰山理论”。 优点：重视创作主体的能动作用，揭示了文学与科学的区别，指出文学不属于认识而属于情感领域。 缺点：①把主观与客观完全对立起来，割断情感与外部世界的联系，导致创作中情感源泉的枯竭。 ②把情感与认识对立起来，割断了情感与理智的联系，导致创作中一味的情 感的宣泄。 ③把表现与规范对立起来，把表现等同于天才、灵感。 2.文学审美的特殊性 文学审美的特殊性表现在： 1)整体性：表现在两个方面，在文学中，生活是一个有机的整体；在文学中，人是一个活 生生的生命整体。例如《杜十娘怒沉百宝箱》，以外写内的文学手段，通过人物行动、外貌等表现人物的心理，杜十娘的出身青楼，对爱情的渴望，以及遇人不淑的事件决定了杜十娘必死的结局。又如在《红楼梦》中，林黛玉身在封建社会，其地位与性情以及过分看重爱情，注定她的宿命。在列夫·托尔斯泰的《安娜·卡列尼娜》中安娜未得到爱情暗示了她的死亡。 2)形象性：事物的外形是怎样的，抓住外在具体的东西。可以以外写内，例写“愁”，将 情感化具体形象，切忌将情感抽象化，或长篇大论，应找具体“意象”，融入主体感情。 3)情感性：文学以人为中心，人是有思想有感情的。而且文学形象带有主体创作色彩，文 学是有情感的价值趋向与判断，没有纯客观的形象。例如“一切景语皆情语”。审美主体将情感移情到审美客体上。 4)无功利性：文学有求真、从善的功能，最终为了求美，这里的美指的是崇高的美，壮美。 例如悲剧，美在其悲剧性，表现主人公的抗争性。文学不追求直接的功利性，而是追求情感的体验，例如歌德诗体小说《少年维特之烦恼》，艺术影响很大，艺术追求积极健
小学数学图形与几何知识点归纳汇总 图形与几何 一线和角 （1）线 *直线 直线没有端点；长度无限；过一点可以画无数条，过两点只能画一条直线。 *射线 射线只有一个端点；长度无限。 *线段 线段有两个端点，它是直线的一部分；长度有限；两点的连线中，线段为最短。 *平行线 在同一平面内，不相交的两条直线叫做平行线。 两条平行线之间的垂线长度都相等。 *垂线 两条直线相交成直角时，这两条直线叫做互相垂直，其中一条直线叫做另一条直线的垂线,相交的点叫做垂足。 从直线外一点到这条直线所画的垂线的长叫做这点到直线的距离。 （2）角 （1）从一点引出两条射线，所组成的图形叫做角。这个点叫做角的顶点，这两条射线叫做角的边。 （2）角的分类 锐角：小于90°的角叫做锐角。 直角：等于90°的角叫做直角。 钝角：大于90°而小于180°的角叫做钝角。 平角：角的两边成一条直线，这时所组成的角叫做平角。平角180°。 周角：角的一边旋转一周，与另一边重合。周角是360°。
1长方形 （1）特征 对边相等，4个角都是直角的四边形。有两条对称轴。 （2）计算公式 c=2(a+b)s=ab 2正方形 （1）特征： 四条边都相等，四个角都是直角的四边形。有4条对称轴。 （2）计算公式 c=4a s=a2 3三角形 （1）特征 由三条线段围成的图形。内角和是180度。三角形具有稳定性。三角形有三条高。（2）计算公式 s=ah/2 （3）分类 按角分 锐角三角形：三个角都是锐角。 直角三角形：有一个角是直角。等腰三角形的两个锐角各为45度，它有一条对称轴。钝角三角形：有一个角是钝角。 按边分 不等边三角形：三条边长度不相等。 等腰三角形：有两条边长度相等；两个底角相等；有一条对称轴。 等边三角形：三条边长度都相等；三个内角都是60度；有三条对称轴。
英语语言学试题（1） I. Directions: Read each of the following statements carefully. Decide which one of the four choices best completes the statement and put the letter A, B, C or D in the brackets. (2%×10=20%) 1、As modern linguistics aims to describe and analyze the language people actually use, and not to lay down rules for "correct" linguistic behavior, it is said to be ___. A、prescriptive B、sociolinguistic C、descriptive D、psycholinguistic 2、Of all the speech organs, the ___ is/are the most flexible. A、mouth B、lips C、tongue D、vocal cords 3、The morpheme "vision" in the common word "television" is a(n) ___. A、bound morpheme B、bound form C、inflectional morpheme D、free morpheme 4、A ___ in the embedded clause refers to the introductory word that introduces the embedded clause. A、coordinator B、particle C、preposition D、subordinator 5、"Can I borrow your bike?" _____ "You have a bike." A、is synonymous with B、is inconsistent with C、entails D、presupposes 6、The branch of linguistics that studies how context influences the way speakers interpret sentences is called ___. A、semantics B、pragmatics C、sociolinguistics D、psycholinguistics 7、Grammatical changes may be explained, in part, as analogic changes, which are ___ or generalization. A、elaboration B、simplification C、external borrowing D、internal borrowing 8、___ refers to a marginal language of few lexical items and straightforward grammatical rules, used as a medium of communication. A、Lingua franca B、Creole C、Pidgin D、Standard language 9、Psychologists, neurologists and linguists have concluded that, in addition to the motor area which is responsible for physical articulation of utterances, three areas of the left brain are vital to language, namely, ___ . A、Broca's area, Wernicke's area and the angular gyrus B、Broca's area, Wernicke's area and cerebral cortex C、Broca's area, Wernicke's area and neurons D、Broca's area, Wernicke's area and Exner's area 10、According to Krashen, ___ refers to the gradual and subconscious development of ability in the first language by using it naturally in daily communicative situations. A、learning B、competence C、performance D、acquisition II. Directions: Fill in the blank in each of the following statements with one word, the first letter of which is already given as a clue. Note that you are to fill in One word only, and you are not allowed to change the letter given. (1%×10=10%) 11、Chomsky defines "competence" as the ideal user's k_______ of the rules of his language. 12、The four sounds /p/,/b/,/m/ and /w/have one feature in common, i.e, they are all b______ . 13、M_______ is a branch of grammar which studies the internal structure of words and the rules by which words are formed. 14、A s______ is a structurally independent unit that usually comprises a number of words to form a complete statement, question or command. 15、Synonyms that are mutually substitutable under all circumstances are called c______ synonyms. 16、The illocutionary point of r_____ is to commit the speaker to something's being the case, to the truth of what has been said. 17、Words are created outright to fit some purpose. Such a method of enlarging the vocabulary is known as word c______.
文学概论知识点整理
1.再现论与表现论 再现论：文学是对现实的模仿，具有实践性。亚里士多德是再现论的奠基人。在中国，最早的文学作品是《二言弹歌》，“断竹续竹，飞土逐肉”再现当时人类的生存环境。 优点：①对文学艺术的源泉做出了唯物主义的解释。 ②十分重视文学艺术的认识价值。 ③引导作家深入生活。 缺点：①没有分清艺术与科学的界限， 认为两者的本质都是给人以真 知，只不过科学是以概念来把握 生活，文学以形象来反映生活。 ②艺术创作的目的是为了逼近生 活，以忠于现实，逼肖现实作为 创作于批评的最高标准，忽视作 者对生活的情感体验。 表现论：文学是心灵的表现。如中国文论中意境的范畴，有我之境与无我之境。外国文论应追溯到柏拉图，他的“迷狂说”是表现论的理论基础，
认为文学艺术陷入迷狂。真正对此理论作出贡献的是康德，认为知、情、意三者同样重要，作者主要表现情感。康德的表现论是是有审美标准的，理性的。此外还有弗洛伊德的“冰山理论”。 优点：重视创作主体的能动作用，揭示 了文学与科学的区别，指出文学 不属于认识而属于情感领域。 缺点：①把主观与客观完全对立起来， 割断情感与外部世界的联系，导 致创作中情感源泉的枯竭。 ②把情感与认识对立起来，割断 了情感与理智的联系，导致创作 中一味的情感的宣泄。 ③把表现与规范对立起来，把表 现等同于天才、灵感。 2.文学审美的特殊性 文学审美的特殊性表现在： 1)整体性：表现在两个方面，在文学中，生活是 一个有机的整体；在文学中，人是一个活生生的生命整体。例如《杜十娘怒沉百宝箱》，以外写内的文学手段，通过人物行动、外貌等表
练习1 1. There is no logical connection between meaning and sounds. A dog might be a pig if only the first person or group of persons had used it for a pig. This is one of the design features of language.A. duality B. arbitrariness C. productivity D. displacement 2. Language is a system of two sets of structures, one of sounds and the other of meaning. This is . It makes people possible to talk everything within his knowledge. A. duality B. arbitrariness C. productivity D. displacement 3. ___ refers to the ability to construct and understand an indefinitely large number of sentences in one’s native language, including those that he has never heard before, but that are appropriate to the speaking situation .A. duality B. arbitrariness C. productivity D. displacement 4. __ __ refers to the fact that one can talk about things that are not present, as easily as he does things present. The dog couldn’t be bow-wowing sorrowfully for some lost love or a bone to be lost. A. duality B. arbitrariness C. productivity D. displacement 5. ______ means language is not biologically transmitted from generation to generation, but the linguistic system must be learnt anew by each speaker. A. duality B. Arbitrariness C. interchangeability D. cultural transmission 6. ______ means that any human being can be both a producer and a receiver of messages. A. duality B. Arbitrariness C. interchangeability D. cultural transmission 7. To say “How are you.” “Hi” to your friends is the ____ __of language. A. directive function B. informative function C. phatic function D. interrogative function 8. “Tell me the result when you finish.” If you want to get your hearer to do something, y ou should use the _____ of language. A. directive function B. informative function C. phatic function D. interrogative function 9. A linguist regards the changes in language and language use as __ ___. A. unnatural B. something to be feared C. natural D. abnormal 10. A linguist is interested in ___A. speech sounds only B. all sounds C. vowels only 11. Which of the following sounds is a voiceless bilabial stop? A. [t] B. [m] C. [b] D. [p 12. Which of the following sounds is a voiced affricate? A. [y] B. [t∫] C. [z] D. [dЗ] 13. Which of the following sounds is a central vowel? A. [ ? ] B. [ i ] C. [ou] D. [a: ] 14. In the following sounds , ______ is a palatal fricative ? A. [ s ] B. [∫] C. [ l ] D. [θ] 15. In the following sounds, _____ is a voiceless affricative? A. [dЗ] B. [v] C. [t∫] D. [θ] 16. In English if a word begins with a [ l ] or [ r ],then the next sound must be a __ __. A. fricative B. nasal sound C. semi-vowel D. vowel 17. Of the “words” listed below___ is not an English word A. [r∧b ] B. [ l? b ] C. [m?sta:∫] D. [lm?p] 18. ___ are produced when the obstruction created by the speech organs is total and audibly released. A. Back vowels B. Stops C. Fricatives D. Glides 19. The International Phonetic Association devised the INTERNATIONAL PHONETIC ALPHABET in _____. A. 1965 B. 1957 C. 1888 D. 1788 20. ___ is a phonological unit , and it is a unit that is of distinctive value. A. Phone B. Phoneme C. Allophone D. Sound 1. [ f ] is a dental consonant. F 2. Phonology studies the characteristics of speech sounds and provides methods for their description, classification and transcription. F 7. The three / p / are allophones. T 3. Phoneme is a phonological unit. T 4. Phone is a phonetic unit. T
Chapter 6 pragmatics 语用学 知识点： 1.*Definition: pragmatics; context 2.*sentence meaning vs utterance meaning 3.*Austin’s model of speech act theory 4.Searle’s classification of speech acts 5.*Grice’s Cooperative Principle 考核目标： 识记：*Definition: pragmatics; context 领会：Searle’s classification of speech acts 综合应用：sentence meaning vs utterance meaning；Austin’s model of speech act theory；Grice’s Cooperative Principle 一、定义 1. Pragmatics语用学: Pragmatics: the study of how speakers of a language use sentences to effect successful communication. Pragmatic can also be regarded as a kind of meaning study.语用学研究的是语言使用者是如何使用句子成功进行交际的。语用学也可以看作是一中意义研究。(它不是孤立地去研究语义，而是把语义置于使用语境中去研究的一门学科。) 2. Context 语境：The notion of context is essential to the pragmatic study of language, it’s generally considered as constituted by the knowledge shared by the speaker and the hearer. 语境这个概念对语言的语用研究来说是必不可少的。一般认为他是由言者和听者的共享知识所构成的。 二、知识点 6.1.2 pragmatics vs. semantics语用学与语义学 二十世纪初，Saussure’s Course in General Linguistics 一书的出版标志着现代语言学研究的开始，同时也为现代语言学奠定了基础调，即语言应该作为一个独立的，内在的系统来加以研究。 语用学和语义学既有相关性又有相异性。两者都是对意义的研究。传统语义学把语义看成是抽象的，内在的，是语言本身的特性，不受语境的影响。因此传统语义学只研究语义的内在特征，不把语义研究置于语境中来考察。语用学研究的是交际过程中语言意义的表达和理解。语用学家认为不把意义放在语境中来考虑就不可能对语义进行充分的描述，因此在研究语义时是否考虑语境便成了传统语义学和语用学的根本区别所在。 Semantics 和Pragmatics的区分 Pragmatics studies how meaning is conveyed in the process of communication. The basic difference between them is that pragmatics considers meaning in context, traditional semantics studies meaning in isolation from the context of use.
有理数数轴：数轴是规定了原点、正方向、单位长度的一条直线。 有理数 概念：凡能写成形式的数，都是有理数。（正整数、0、负整数统称整数；正分数、负分数统称分数；整数和分数统称有理数.注意：0即不是正数， 也不是负数；-a不一定是负数，+a也不一定是正数；π不是有理数。） 有理数的分类：①有理数 正有理数 零 负有理数 正整数 正分数 负整数 负分数 ②有理数 整数 分数 正整数 零 负整数 正分数 负分数 相反数 (1)只有符号不同的两个数，我们说其中一个是另一个的相反数；0的相反数还是0； (2)相反数的和为0 ? a+b=0 ? a、b互为相反数。 绝对值：正数的绝对值是其本身，0的绝对值是0，负数的绝对值是它的相反数；注意：绝对值的意义是数轴上表示某数的点离开原点的距离； 有理数比大小 （1）正数的绝对值越大，这个数越大； （2）正数永远比0大，负数永远比0小； （3）正数大于一切负数； （4）两个负数比大小，绝对值大的反而小； （5）数轴上的两个数，右边的数总比左边的数大； （6）大数-小数＞0，小数-大数＜0。 互为倒数：乘积为1的两个数互为倒数；注意：0没有倒数；若a≠0，那么的倒数是； 若ab=1? a、b互为倒数；若ab=-1? a、b互为负倒数。 有理数乘方的法则 （1）正数的任何次幂都是正数； （2）负数的奇次幂是负数；负数的偶次幂是正数；注意：当n为正奇数时: (-a)n=-a n 或(a -b)n=-(b-a)n , 当n为正偶数时: (-a)n =a n 或(a-b)n=(b-a)n . 科学记数法：把一个大于10的数记成a×10n的形式，其中a是整数数位只有一位的数，这种记数法叫科学记数法. 近似数的精确位：一个近似数，四舍五入到那一位，就说这个近似数的精确到那一位。 有效数字：从左边第一个不为零的数字起，到精确的位数止，所有数字，都叫这个近似数的有效数字。 举几个例子：3一共有1个有效数字，0.0003有一个有效数字，0.1500有4个有效数字， 1.9*10^3有两个有效数字（不要被10^3迷惑，只需要看1.9的有效数字就可以了，10^n 看作是一个单位）。 整式的加单项式：在代数式中，若只含有乘法（包括乘方）运算。或虽含有除法运算，但除式中不含字母的一类代数式叫单项式。 单项式的系数与次数：单项式中不为零的数字因数，叫单项式的数字系数，简称单项式的系 数；系数不为零时，单项式中所有字母指数的和，叫单项式的次数。 多项式：几个单项式的和叫多项式。
语言学概论试题及答案 分享 首次分享者：◇﹎ゞ丫丫℡已被分享11次评论(0)复制链接分享转载举报语言学概论形成性考核作业及参考答案 语言学概论作业1 导言、第一章、第二章 一、名词解释 1、历时语言学——就各种语言的历史事实用比较的方法去研究它的“亲属”关系和历史发展的，叫历时语言学。 2、语言——语言是一种社会现象，是人类最重要的交际工具和进行思维的工具。就语言本身的结构来说，语言是由词汇和语法构成的系统。 3、符号——符号是用来代表事物的一种形式，词这样的符号是声音和意义相结合的统一体。任何符号都是由声音和意义两方面构成的。 4、语言的二层性——语言是一种分层装置，其底层是一套音位；上层是音义结合的符号和符号的序列，这一层又分为若干级，第一级是语素，第二级是由语素构成的词，第三级是由词构成的句子。 5、社会现象——语言是一种社会现象和人类社会有紧密的联系。所谓“社会”，就是指生活在一个共同的地域中，说同一种语言，有共同的风俗习惯和文化传统的人类共同体。语言对于社会全体成员来说是统一的、共同的；另一方面，语言在人们的使用中可以有不同的变异、不同的风格。 二、填空 1、结构主义语言学包括布拉格学派、哥本哈根学派、美国描写语言学三个学派。 2、历史比较语言学是在19世纪逐步发展和完善的，它是语言学走上独立发展道路的标志。 3、人的大脑分左右两半球，大脑的左半球控制语言活动，右半球掌管不需要语言的感性直观思维。 4、一个符号，如果没有意义，就失去了存在的必要，如果没有声音，我们就无法感知，符号也就失去了存在的物质基础。 5、用什么样的语音形式代表什么样的意义，完全是由使用这种语言的社会成员约定俗成。 6、语言符号具有任意性和线条性特点。 7、语言的底层是一套音位，上层是符号和符号的序列，可以分为若干级，第一级是语素，第二级是词，第三级是句子。 8、语言系统中的所有符号，既可以同别的符号组合，又可以被别的符号替换，符号之间的这两种关系是组合和聚合。 9、组合是指符号与符号相互之间在功能上的联系，聚合是指符号在性质上的归类。 三、判断正误（正确的打钩，错误的打叉） 1、文字是人类最重要的交际工具。（×） 2、地主阶级和农民阶级之间没有共同语言，这说明语言是有阶级性的。（×） 3、在现代社会，文字比语言更加重要。（×）
文学概论期末复习重点 1.文艺学是一门以文学为对象，以揭示文学基本规律，介绍相关知识为目的的学科，包括三个分支，即文学理论、文学批评和文学史。 2.文学四要素：作品、作家、世界、读者 3.文学理论所规定的五个方面的任务：（1）文学活动发展论：文学活动作为人类的一种精神活动，随着时代的发展而发展；（2）文学活动本质论：文学作为人类的一种特殊的精神活动，从总体上来研究文学活动区别于其他活动的特殊性质；（3）文学创作论：“世界”就是我们所指的社会生活，研究作家如何根据生活进行艺术创造的过程和规律；（4）作品构成论：作家创作出来的文学文本在阅读、研究和批评中变成了作品，研究作品的构成因素及其相互关系；（5）文学接受论：研究读者接受过程和规律。 4.模仿论与表现论的区别：（1）模仿论认定文学是世界的反映，表现论则认定文学是作者心灵的表现。(2)表现论强调作者对作品意义的生成作用，模仿论更强调了解作品所描写的世界和写作背景。（3）表现论将文学创作同科学研究等活动对立起来，模仿论强调文学创作应遵循的客观规律。（4）（可略）模仿论：文学活动是一种意识反应行为。立论观点是“文艺起源于人对宇宙或世界的模仿”。柏拉图从模仿世界的观点出发，在结论上否定了艺术存在的合理性；亚里士多德却认为，艺术的起源和主要作用在于模仿，它不仅反映事物的外观形态，而且反映事物的内在规律和本质。 表现论：作者通过创作表达感受并以此唤起读者相近的感受。 5.文学起源说：（1）巫术发生说：弗雷泽，“相似律”同类相生或同果必同因，“接触律”甲乙二物接触后，施力于甲可影响乙，施力于乙可影响甲；田野调查（2）宗教发生说：赫尔德“灵魂的音乐”，格罗塞（3）游戏发生说：康德“其本质特征就是无目的的合目的性或自由的合目的性。（4）劳动说：马克思，劳动提供了文学活动的前提条件，劳动产生了文学活动的需要，劳动构成了文学描写的主要内容，劳动制约了早期文学的形式。 6.文学的审美意识形态属性的表现： （1）无功利与功利： A.文学是无功利的，但这种无功利本身也隐含有某种功利意图。 B.文学的这种无功利性集中体现在作家的创作活动和读者的阅读过程中。 C.作为作家或读者的话语活动，文学虽然与直接的功利目的无关，但简洁地仍旧有深刻的社会功利性。作为再现现实社会生活的话语结构，文学的功利性在于，它把审美无功利性仅仅当做实现其再现社会生活这一功利目的的特殊手段。（2）形象与理性 文学是形象的，这是由文学直接的审美感性特征决定的；同时，文学又蕴含某种理性，即文学创作、阅读及形象本身都可能与某种间接的或深层的理性考虑有关，这是由文学的人类活动属性本身决定的。 （3）情感与认识 审美情感是直接的而理智认识是间接的。直接的审美情感在其深层往往隐伏着间接的理智认识。《我爱这土地》 7.话语是与语言、语言系统、言语和文本等存在联系和区别的概念。语言是人类最重要的社会交际工具，话语则是具体的社会存在形态。