下载文档原格式
J.Kegl’s paper,‘‘Machine Readable Dictionaries and Educa-tion:How people are able to talk about what they see.Its chap-tion’’has been updated and revised.Focusing on primary educa-ters derive from a conference on language and space that was tion,it provides an overview of the problem area,and updates held in Tucson in1994.The book is,however,more than a research efforts which purportedly resulted from the Workshop.conference proceedings,because the individual papers that form It lives up to the stated thesis of the collection,to not only recap the basis of each chapter have been edited to incorporate the the Workshop but to discuss how thefield has progressed and interaction that occurred during the conference to create a more what questions remain prehensive overview of the linguistics of spatial discourse.‘‘Why Use Words to Label Ideas:The Uses of Dictionaries The topic with which the authors are dealing can be under-and Thesauri in Information Retrieval’’is an excellent overview stood intuitively by considering two ways in which humans by M.Lesk.Lesk discusses how one might use a dictionary as a interact with their environment.Thefirst way is by dealing with thesaurus for information retrieval,and the problems and issues things in space/time.People perceive,react to,and manipulate involved including the issue of multiword phrases in retrieval;external objects.Our brains are equipped with facilities for and the distinction between lexicographers who‘‘split’’differ-dealing with spatio-temporal locations,and the aspects of reality ences and thesauri constructors who‘‘lump’’similarities.Nev-that may be found in those locations.Through these spatial ertheless,to include a section on Automatic Indexing and not representation facilities,people learn about and understand real-include references past the Workshop does not do justice to ity.The second way humans interact with their environment ‘‘seeing how things change;and how they stay the same.’’is through language.They perceive language,interpret it,and In‘‘Structure and Access in an Automated Lexicon and understand it.Our brains are equipped with facilities for dealing Related Issues,’’N.Calzolari elaborates the differences between with conceptual structures as expressed in propositions,state-the more generalized,machine-readable dictionary(MRD),and ments,questions,and other forms of linguistic and symbolic what Calzolari calls a‘‘structured lexical database’’which in-expression.People learn about and understand reality through cludes lexical facts not generally found in an MRD.While this these conceptual facilities,as well as through spatial representa-paper has not been updated,it functions well,adding another tion.foundational layer to the collection.Learning and understanding occur in parallel ways through To recap,the stated goal of the book is to provide a‘‘baselinethese two different cognitive facilities:The spatial representa-and reference point for further research and development’’intion facilities and the conceptual facilities of the brain.A simple thefield and to‘‘organize research’’and‘‘further co-opera-example may illustrate the way the same outcome can occur tion’’;in general these goals have been met.The weakest partthrough these two different kinds of processing.A person can of this collection is that the papers are uneven in quality andlearn that it is raining outside by going(or looking)outside timeliness and much of the baseline material is repeated,whichand experiencing rain.Or that person can hear from someone may have been unavoidable,but it nevertheless becomes pro-(whose information is considered trustworthy)that it is raining. saic.Like other aspects of thefield,automated lexicon buildingIn one case,the spatial representation system of the brain is has not remained static,thus it is curious that a paper aboutused to go outside and observe the phenomenon.In the other computational linguistics research,‘‘Machine-Readable Dic-case,the conceptual structures of the brain are used to under-tionaries and Computational Linguistics Research,’’by B.K.stand the message conveyed.This simple example also shows Boguraev,would not have been updated since1987.Other pa-the necessity of interactions between the two cognitive systems. pers,most notably the one written by Kegl not only containsTo understand the message that it is raining outside,the spatial contemporary,distinctive information,but reports on her ownrepresentation for‘‘outside’’must be employed by the concep-research since the Workshop.tual system.Again,the editors motivation in publishing a collection ofThe discussion in this book deals with one aspect of the papers which were presented at the Marina di Grosetto Work-interaction between language and space:The act of talking about shop has been sustained.Creating something other than a com-space,which requires using elements of conceptual structures to pilation of disparate papers from a Workshop is a difficult task,represent spatial relationships.For example,language contains but it has been satisfactorily accomplished.The collection isprepositions that present spatial relationships such as‘‘above,’’readable,and perhaps because it is a publication of Workshop‘‘below,’’‘‘next to,’’‘‘behind,’’etc.It also contains nouns that talks,almost breezy in its tone.The benefit of this is that itpresent spatial locations such as‘‘top,’’‘‘side,’’and‘‘front,’’makes difficult concepts accessible to the non-expert in thefield.and adjectives that present spatial dimensions such as‘‘high’’This collection of Workshop papers will be useful for a graduateor‘‘wide.’’Using language,speakers adopt different frames of seminar course which covers computational linguistics,particu-larly outside the domain.While professionals in thefield will reference,and different levels or types of spatial reference. already be familiar with most of what has been assembled here,The study of speech about spatial representations leads to an as an introduction to thefield,this collection will hold up well,understanding of how the spatial facilities of the brain and its if unevenly,for another decade to come.conceptual structures are linked.The chapters by Jackendoff;Bierwisch;and Peterson,Nadel,Bloom,and Garrett serve assolid introductions to this special branch of linguistics.P.Zoe¨Stavri The relationship between conceptual and spatial facilities isResearch and Education a complex one.Consider the examples of the spatial languageNational Library of Medicine associated with the preposition‘‘in’’given by Bierwisch in hisBethesda,MD20894chapter.When we say,‘‘Thefish is in the water,’’we meanE-mail:**************.govsomething rather different than when we say,‘‘The boat is inthe water.’’When we say,‘‘He has a strawberry in his mouth,’’we mean something rather different that when we say,‘‘He hasa pipe in his mouth.’’When we say,‘‘There are some coins in Language and Space.Paul Bloom,Mary A.Peterson,Lynnthe purse,’’we mean something rather different than when we Nadel,and Merrill F.Garrett(eds.).Cambridge,MA:The MITsay,‘‘There is a hole in the purse.’’And then there is the rather Press;1996:597pp.Price:$50.00.(ISBN0-262-02403-9.)different range of meanings that can be associated with‘‘He isnot in today.’’The chapter by O’Keefe presents an interesting This book presents a good overview of one aspect of theinteraction between spatial and conceptual processes in cogni-perspective on such prepositions that goes beyond simple lin-JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE—April1,1998381guistic analysis,associating spatial prepositions with a cognitive retrieval,information science research has recently been investi-map in the brain.gating systems that rely on the interaction of the conceptual Other phenomena illustrate the complexity of the interaction and spatial systems of the brain.between the conceptual facilities of the brain that process lan-Although the contents of this volume relate to areas of cur-guage and the spatial facilities of the brain that process spatial rent information research,the approach taken by the authors perception.Among these are the problems of frames of refer-tends to obscure the relationship.The narrow focus on the way ence,and of perspective,as discussed by Levelt,Levison,and in which language is used to talk about space precludes abroader discussion of the aspects of the interaction of conceptual Tversky.In speaking about spatial phenomena,it seems neces-sary to adopt a frame of reference.There are several alternative and spatial cognition that are of current interest in information frames of reference available,and these differ from one lan-science.As a result,those who look to this book to provide a guage to another.Similarly,an understanding of the way lan-basis for the design of visualization systems or of interfaces will guage about space develops in language learners(Landau),andbe disappointed.What it does provide is a clear and thorough how neurological deficits can affect this capability(Shallice),introduction to the linguistics of spatial language,and this intro-provide insights into the interaction.The range of investigationsduction will be of value to students and researchers who wish represented in this book is extended by discussions of signed to approach information systems from afirm basis of under-language(Emmorey)and offictive motion(Talmy).Signedstanding cognitive processes and structures.language brings the spatial and conceptual systems into a rela-tionship that seems somewhat different from verbal language,and an analysis offictive motion(the application of motion-Bryce Allenrelated words to things that do not really move,as in sayingSchool of Library and Informational Science that a fence goes from this spot to that spot)provides insightUniversity of Missouriinto the metaphorical value of spatial language.105A Stewart Hall Perhaps the most interesting chapters in this book are thoseColumbia,MO65211that explore the question of the relative priority of one systemE-mail:***********************.eduover the other.Chapters by Mandler and Johnson-Laird suggestthat the basics of meaning and of deductive thinking can befound in spatial representations and in particular in mental mod-els:Conceptual representations in the brain that make use offeatures of spatial representations.This perspective is supportedby the chapter by Logan and Sadler that presents a computa-The Economics of Information:A Guide to Economic and tional analysis of spatial relationships.On the other hand,the Cost-Benefit Analysis for Information Professionals.Bruce chapter by Bowerman points out that non-linguistic representa-R.Kingma.Englewood,CO:Libraries Unlimited;1996:200 tions alone cannot provide the conceptual packaging needed pp.Price:$35.00.(ISBN:1-56308-303-5.)for the expression of spatial relationships.This cross-linguisticperspective suggests that the cognitive systems for spatial repre-Money and all its related scholarlyfields such asfinance sentation and for conceptual processing interact dynamically.and economics have,until recently,been peripheral areas of The theory and research presented in this book is of interest study in LIS education.Tighter budgets,higher costs for capital to those who seek to understand how cognition works.It is also investments,and the need to plan,evaluate,and justify the of interest to those information scientists who are investigating existence of services infinancial terms,however,require infor-topics and issues that require the interaction of the conceptual mation professionals to develop more sophisticated skills in and spatial systems.For example,investigations of the effec-making economic decisions.Although some might argue with tiveness and impact of different types of information representa-the author’s assertion that economics‘‘provides an unbiased tion on information retrieval and learning require an understand-objective method of analysis,’’(p.4)for making such decisions, ing of this interaction.Considerations of how images enhance The Economics of Information is a useful resource for training information in documents and how images can be retrieved LIS professionals.using language are examples of this type of research.The text is divided into three sections that examine a general Other information scientists are concerned with mental mod-introduction to economic principles,the nature of information els,particularly as they facilitate retrieval,and as they can be markets and market failure,and a general introduction to cost-embodied in interface designs.These mental models may in-benefit analysis,respectively.The prose style is clear,straight-clude spatial representations that structure and constrain the forward,and largely free of unexplained technical jargon.Each conceptual tasks of information retrieval.Examples of inter-chapter in the book is generously supplied with examples from faces based on spatial mental models include desktop and the context of information markets that make them particularly ‘‘room’’metaphors.As the chapter by Johnson-Laird demon-useful for classroom use or as a primer for practitioners and strates,mental models are formed by the interaction of spatial researchers who are interested in the application of economics and conceptual representations.The discussion of mental mod-to information use.els in this book is,however,at a conceptual rather than an As the author correctly notes,many of the economic con-applied level,and relates to deductive thinking and temporal cepts used in the book are not new in and of themselves.Given reasoning rather than to aspects of seeking and manipulating that many LIS students have limited exposure to economics information.andfinance,some general background is probably necessary to Afinal area of great current interest is that of visualization.discuss specific information-related areas.Readers who have Here,information spaces,which are largely conceptual in na-a background in economics orfinance mayfind themselves ture,are represented spatially.In other words,spatial perception skimming chapters tofind specific applications for information is applied,at least metaphorically,to conceptual structures.One markets.However,even these readers are likely tofind enough can skim over ideas,fold one idea into another,or visualize a new applications of economics concepts to make the book a landscape of concepts.Whether it is the topographical mapping useful introduction to the unique problems that involve valuing of scientific domains produced by co-citation analysis,the pre-information.The second section of the text is recommended for sentation of information for two-or three-dimensional brows-its discussion of information markets and pricing.ing,or the spatial representation of the output of informationTo the extent that the book has a weakness,it is in the third 382JOURNAL OF THE AMERICAN SOCIETY FOR INFORMATION SCIENCE—April1,1998。
0 why do people say that Chinese are the offspring of dragons？According to one Chinese fairy tale, the first dragon’s name was Fuxi. Fuxi had a dragon’s body and man’s head. He married his sister, Nuwa, who then give birth to the first group of men and women on earth. The Yellow Emperor also had a dragon’s body and man’s head, and so was named the Yellow Dragon. The Yellow Emperor’s brother, Emperor Yan, also had the body of dragon. Some say the Chinese people are descendants of the Yellow Emperor and Emperor Yan and re thus descendants of dragon.1 four cooking schools in China(1) The Cantonese DishesA. fresh materials and great tendernessB. fresh water fish, seafood, birdsC. oyster, sauce, fermented soy bean sauce, fish sauce, lard sugar and vinegarD. stir-frying, steaming(2) The Sichuan DishesA. distinct and variousB. wild edible herbs, and the meat of domestic animals and birdsC. fish flavors, pepper powder boiled in oil, strange flavor and sticky-hotD. sauting, stir-frying without stewing, dry braising(3) The Shandong DishesA. domestic animals and birds, seafood and vegetablesB. BAO, LIU, PA roasting and boiling, using sugar to make fruitC. tasty and freshD. wide selection of materials, cooking methods, and seafood(4) The Huaiyang DishesA. cleanliness and freshness of its ingredientsB. natural juices and flavorsC. lightness, freshness, sweetness, and mildness of tasteD. explode-frying, red-stew, stir-frying2火锅the hot pot（译文）源于中国北部的火锅有1000多年的历史。
丝绸之路的英文演讲稿The Silk Road, also known as the Silk Route, was a network of trade routes that connected the East and the West, spanning from the ancient Chinese capital of Chang'an all the way to the Mediterranean Sea. This historic route was not only a pathway for the exchange of goods, but also a bridge for cultural and technological exchanges between the East and the West.The Silk Road was named after the lucrative trade in silk that was carried out along its routes. However, it was not just silk that was traded; a wide variety of goods such as spices, precious metals, and exotic animals were also exchanged. This exchange of goods not only enriched the economies of the regions along the route, but also facilitated the spread of ideas, religions, and technologies.One of the most significant impacts of the Silk Road was the cultural exchange that took place along its routes. The interaction between different civilizations led to the transmission of artistic, architectural, and philosophical ideas. For example, the spread of Buddhism from India to China and the introduction of Greek and Roman art and architecture to the East are clear evidence of the cultural diffusion that occurred along the Silk Road.Furthermore, the Silk Road played a crucial role in the exchange of knowledge and technology. The transfer of scientific and technological knowledge, such as the production of paper, the compass, and gunpowder, contributed to the advancement of civilizations along the route. The exchange of medical knowledge and practices also had a profound impact on the development of traditional medicine in different regions.In addition to the economic, cultural, and technological exchanges, the Silk Road also facilitated the exchange of religions. Buddhism, Islam, Christianity, and Zoroastrianism were among the religions that spread along the route, leading to the diversification of religious beliefs and practices in different regions.The legacy of the Silk Road is still evident today. The cultural diversity and interconnectedness of the regions along the historic route continue to influence the modern world. The concept of the Silk Road has also been revitalized in the form of the Belt and Road Initiative, a global development strategy adopted by the Chinese government to promote economic cooperation and connectivity among countries along the ancient Silk Road routes.In conclusion, the Silk Road was not just a trade route, but a conduit for the exchange of goods, ideas, and cultures. Its impact on the development of civilizations along its routes was profound and far-reaching. The legacy of the Silk Road continues to resonate in the modern world, serving as a reminder of the importance of connectivity and exchange in shaping the course of human history. Thank you.。
化学镀Ni -Cu -P 工艺天津大学材料系(300072) 崔振铎 杨贤金 王 慧 邓才君 岳松山摘 要 研究了不同Cu 2+浓度、pH 值及温度条件下的化学镀N i-Cu-P 工艺。
结果表明,提高化学镀液温度或在较高的Cu 2+浓度下,可加快化学镀速度,提高镀层中Cu 含量。
镀层在较高Cu 量下存在的少量化合物N i 5P 2能使沉积颗粒细化。
含Cu 镀层具有较高硬度。
主题词 化学镀 合金镀层 耐蚀性1 前 言化学镀N i -P 镀层是N i -P 合金的无定形结构,具有较高的硬度、耐蚀性、耐磨性,因而应用较广。
由于近年来镀层的需求多样化,改善其合金化特性的研究也越来越受到人们的重视[1],如N i-W-P 、N i-M o -P 、Fe -N i-P 等多种镀层。
又因N i-Cu-P 化学镀层具有优良的耐蚀性、导电性和热稳定性[2]而引起了人们的高度重视。
目前的报道多为较低Cu 2+镀液条件下的N i-Cu-P 化学镀工艺及镀层性能,而对较高Cu 2+浓度下的研究却较少。
为此,本文对较高Cu 2+浓度下N i-Cu-P 合金化学镀工艺、镀层结构及性能进行了研究。
2 实 验2.1 镀液成分镀液成分见表1。
表1 化学镀液成分(g /L )镀液CuS O 4・5H 2ONiS O 4・7H 2ONaH 2PO 2・H 2ONa 3C 6H 5O 7・2H 2ONaAc NH 4Ac PbS A 0.22820300200B 0.52820300200C2020202×10-3 C 镀液是酸性N i -P 化学镀液,pH 值为5.5,用浓硫酸调整。
A 、B 为碱性N i -Cu -P 化学镀液,pH 值用氨水调至7.6～8.5。
2.2 实验内容化学镀在水温为37～100℃的电热水浴中进行,采用X -650型扫描电镜观察镀层表面形貌,以ED AX 能谱仪测定镀层中N i 、P 、Cu 含量,用-3型显微硬度计测定镀层硬度。
第一章希腊罗马文化第一节希腊文化Greek Cultur1. Which culture reache e d a high point of development in the 5th century B.C.?A.Greek CultureB.Roman CultureC.Egyptian CultureD.Chinese Culture2. In( )the Romans conquered Greece.A.1200B.C. B.700B.C.C.146B.C.D.the 5th century3. Which of the following works described the war led by Agamemnon against the city of Troy?A.Oedipus the kingB.IliadC.OdysseyD.Antigone4. Which of the following is NOT the greatest tragic dramatist of ancient Greece?A.AristophanesB.EuripidesC.SophoclesD.Aeschylus5. Which of the following is NOT the play written by Aeschylus?A.AntigoneB.AgamemnonC.PersiansD.Prometheus Bound6. Which of the following is NOT the play written by Sophocles?A.ElectraB.AntigoneC.Trojan WomanD.Oedipus the king7. Which of the following is the play written by Euripides?A.AntigoneB.PersiansC.ElectraD.Medea8. Who was the founder of scientific mathematics?A.HeracleitusB.AristotleC.SocratesD.Pythagoras9. Who ever said that "You can not step twice into the same river."?A.PythagorasB.HeracleitusC.DemocritusD.Aristotle10. ( )believed that the highest good in life was pleasure, freedom from pain and emotional upheaval.A.SophistsB.CynicsC.SkepticsD.Epicureans11. ( )is said to have told the king :"Give me a place to stand, and I will move the world."A.ArchimedesB.AristotleC.PlatoD.Euclid第二节罗马文化Roman Culture1. Who wrote "Captive Greece took her rude conqueror captive."?A.SapphoB.PlatoC.VirgilD.Horace2. Increasingly troubled by the inroads of northern tribes such as Goths, the West Roman Empire finally collapsed in( ) A.D..A.395B.27C.1453D.4763. Who wrote "I came ,I saw,Iconquered"?A.HoraceB.Julius CaesarC.VirgilD.Marcus Tullius Cicero4. The author of the philosophical poem On the Nature of Things is( )A.VirgilB.Julius CaesarC.HoraceD.Lucretius5. Which of the following is not Roman architecture?A.The ColosseumB.The PanthenonC.The ParthenonD.Pont du GardTrue or False1. Greek cuiture reached a high point of development in the 6th centurg B.C2. the lliad deals with the alliance of the states of the southern mainlang of Greece,led by Agamemnon in their war against the city of Troy.3. sappho,was considered the most important lyric poet of ancient Greece.4. Herodotus is offen called ;Father of History’ and he wrote about the wars between Greeks and Romans.5. The greatest names in European philosophy are Socrates,piato and Aristotle ,who were active in the 5th and4th century B.C6. Socrates ever said ,;you cannot step twice into the same river’7. Archimedes is ever now well-now for his Elements ,a testbook of geometry.-第二章《圣经》与基督教1. Which of the following is by far the most influential in the West?A.BuddismB.IslamismC.ChristianityD.Judaism2. The Old Testament consists of 39 books, the oldest and most important of which are the first five books, called( )A.ExodusmandmentsC. AmosD.Pentateuch3. At the age of 30, Jesus Christ received the baptism at the hands of _____.A.St.PeterB.St.PaulC.John BaptistD.John Wycliff4. Which of the following emperors issued the Edict of Milan and made Christianity legal in 313? ____.A.Augustus IB.Thedosius IC.Nero ID.Constantine I5. Which of the following emperors made Christianity the official religion of the empire and outlawed all other religions in 392 A.D.?A.TheodosiusB.AugustusC.Constantine ID. Nero Caesar6. By 1963, the whole of the Bible had been translated in _____languages.A.288B.974C.1202D.1547. When printing was invented in the 1500’s, the ____Bible was the first complete work printed.A.EnglishtinC.AramaicD.Hebrew8. When did the standard American edition of the Revised V ersion appear? ____A.1885B.1611C.1901D.1979第三章中世纪第一节庄园与教堂1. In the latter part of the 4th century, which of the following tribles swept into Europe from central Asia, robbing and killing large numbers of the half civilized Germanic tribes? ____A.the MongoliansB.the HunsC.the TurkishD.the Syrians2. The Middle Ages id also called the ____.A. "Age of Christianity"B. "Age of Literature"C. "Age of Holy Spirit"D. "Age of Faith"3. In 732, who gave his soldiers estates known as fiefs as a reward for their service? ___A.Charles Martel, a Frankish rulerB. Charles I, a Turkish rulerC. Constantine I, a Frankish rulerD. St.Benedict, an Italian ruler4. According to the code of chivalry, which of the following is not pledged to do for a knight? _____A.to be loyal to his lordB. B.to fight for the churchC. to obey without question the orders of the abbotD. to respect women of noble birth5. When was the Church divided into the Roman Catholic Church and Eastern Orthodox Church? _____A.after 1066B. after 1296C. after 1054D. after 4766. Under feudalism, what were the three classes of people of western Europe? ____A.clergy, knights and serfsB. Pope, bishop and peasantsC.clergy, lords and peasantsD. knights, nobles and serfs7. By which year the Moslems had taken over the last Christian stronghold and won the crusaders and ruled all the territory in Palestine that crusaders had fought to control? _____A.1270B.1254C.1096D.1291第二节学术、科学、文学艺术与建筑8. Which of the following was crowned "Emperor of the Romans” by the Pope in 800? ____A. St.Thomas AquinasB. CharlemagneC. ConstantineD. King James9. Who was the ruler of the Anglo-Saxon kingdom of Wessex and contributed greatly to the medieval European culture? ____A.Charles IB. Constantine IC. Alfred the GreatD. Charles the Great10. Which country’s epic does Song of Roland belong to? ____A.EnglishB. GermanicC. HebrewD.French。
Revision Contents:Unit 1 Language and Learning1. What are the major views of language? What are their implications to language teaching or learning?Structural View: It sees language as a linguistic system made up of various subsystem:from phonological, morphological, lexical, etc. to sentence. Each language has a finite number of such structural items.To learn a language means to learn these structural items so as to be able to understand and produce language.Audiolingual approach: The teaching of a second language through imitation, repetition, and reinforcement. It emphasizes the teaching of speaking and listening beforereading and writing and the use of mother tongue in the classroom is not allowed. Theprincipal features of audiolingualism are an emphasis on structures in the language whichcan be learned as regular patterns of verbal behavior and the belief that learning is a process of habit formation.Functional View: It sees language as a linguistic system but also as a means for doingthings. Most of our day-to- day language use involves functional activities: offering, suggesting, advising, apologizing, etc. Therefore, learners learn a language in order to dothings with it. To perform functions, learners need to know how to combine thegrammatical rules and the vocabulary to express notions that perform the functions. Communicative approaches are based on this view of language.Interactional View: It considers language as a communicative tool, whose main use is tobuild up and maintain social relations between people. Therefore, learners not only needto know the grammar and vocabulary of the language, but also need to know the rules forusing them in a whole range of communicative context.Some of the language learning approaches and methods based on this view of language are: Strategic interaction; communicative approaches.2. What are the major Views on language learning? What are their implications to language teaching?Behaviouralist theoryBased on the theory of conditioning, Skinner suggested language is also a form of behaviour. It can be learned the same way as an animal is trained to respond to stimuli.This theory of language learning is referred to as behaviouralism, which was adopted forsome time by the language teaching profession, particularly in America.One influential result is the audio-lingual method, which involves endless “listen and repeat” drilling activities. The idea of this method is that language is learned by constantrepetition and the reinforcement of the teacher. Mistakes were immediately corrected, andcorrect utterances were immediately praised. This method is still used in many parts of theworld today.Cognitive theoryIt seems to be largely the result of Noam Chomsky’s reaction to Skinner’s behavioural theory, which led to the revival of structural linguistics.The key point of Chomsky’s theory is reflected in his most famous question: if language is a learned behaviour, how can a child produce a sentence that has never beensaid by others before.One influential idea is that students should be allowed to create their own sentencesbased on their understanding of certain rules. This idea is clearly in opposition to the Audio-Lingual Method.According to the cognitive theory, learning is a process in which the learner activelytries to make sense of data. The basic technique associated with a cognitive theory of language learning is the problem-solving task.Constructivist theory: Learning is a complex cognitive process in which the learner constructs meaning based on his or her own experiences and what he /she already knows.Implications for classroom teachingTeaching should be built based on what learners already know and engage learnersin learning activities.It is believed that education is used to develop the mind, not just to rotate or recallwhat is learned.Teachers need to design activities to interact with learners to foster inventive, creative,critical learners.Teachers must balance an understanding of the habits, characteristics as well as personalities of individual learners with an understanding of the means of arousing learners’ interest and curiosity for learning.Socio-constructivist theory: It emphasizes interaction and engagement with the target language in a social context based on the concept of ‘Zone of Proximal Development’(ZPD) and scaffolding.Learning is best achieved through the dynamic interaction between the teacher andthe learner and between learners. With the teacher’s scaffolding through questions and explanations, or with a more capable peers’ support, the learner can move to a higher level of understanding and extend his / her skills and knowledge to the fullest potential.Unit 2 Communicative Principles and Activities1. The goal of CLT is to develop students’communicative competence.2.What is communicative compentence? Try to list some of its components and theirimplication to teaching.Communicative compentence refers to both the knowledge about the language and the knowledge about how to use the language appropriately in communicative situations. According to Hedge, it includes five components.Linguistic competence --- knowledge of the language itself, its form and meaningPragmatic competence --- the appropriate use of language in social contextDiscourse competence --- one’s ability to create coherent written text or conversation andthe ability to understand them (ability to express or to understand a topic logically and coherently by effectively employing or comprehending the cohesive markers used in the discourse /ability to initiate, develop, enter, interrupt, check, or confirm in a conversation) Strategic competence--- strategies one employs when there is communication breakdown due to lack of resourcesFluency---- one’s ability to ‘link units of speech together with facility and without strain or inappropriate slowness or undue hesitationImplications for teaching and learning:Linguistic competenceTeachers need to help learners----achieve accuracy in the grammatical forms of the language;----pronounce the forms accurately;----use stress, rhythm, and intonation to express meaning;----build a range of vocabulary;----learn the script and spelling rules;----achieve accuracy in syntax and word formation.Pragmatic competenceTeachers need to help learners---learn the relationship between grammatical forms and functions;---use stress and intonation to express attitude and emotion;---learn the scale of formality;---understand and use emotive tone;---use the grammatical rules of language;---select language forms appropriate to topic, listener, or setting, etc.Discourse competenceTeachers need to help learners----take longer turns, use discourse markers and open and close conversations;----appreciate and be able to produce contextualised written texts in a variety of genres;----be able to use cohesive devices in reading and writing texts;----be able to cope with authentic texts.Strategic competenceTeachers need to enable learners----to take risks in using the language;----to use a range of communicative strategies;----to learn the language needed to engage in some of these strategies, e.g. ‘What do you call a thing that/person who…’FluencyTeachers need to help learners-----deal with the information gap of real discourse;-----process language and respond appropriately with a degree of ease;-----be able to respond with reasonable speed in ‘real time”.3.What is communicative language teaching?Communicative language teaching began in Britain in the 1960s as a replacement to。
a rXiv:as tr o-ph/1198v14Nov2The Interaction between the ISM and Star Formation in the Dwarf Starburst Galaxy NGC 4214Fabian Walter California Institute of Technology Christopher L.Taylor Five College Radio Astronomy Observatory Susanne H¨u ttemeister Radioastronomisches Institut,Universit¨a t Bonn Nick Scoville California Institute of Technology Vincent McIntyre Australia Telescope National Facility,CSIRO ReceivedABSTRACTWe present thefirst interferometric study of the molecular gas in the metal–poor dwarf starburst galaxy NGC4214.Our map of the12CO(1–0)emission, obtained at the OVRO millimeter array,reveals an unexpected structural wealth. We detected three regions of molecular emission in the north–west(NW),south–east(SE)and centre of NGC4214which are in very different and distinct evo-lutionary stages(total molecular mass:∼5.1×106M⊙).These differences are apparent most dramatically when the CO morphologies are compared to optical ground based and HST imaging:massive star formation has not started yet in the NW region;the well–known starburst in the centre is the most evolved and star formation in the SE complex started more recently.We derive a star formation efficiency of∼8%for the SE ing high–resolution VLA observations of neutral hydrogen(H I)and our CO data we generated a total gas column density map for NGC4214(H I+H2).No clear correlation is seen between the peaks of H I,CO and the sites of ongoing star formation.This emphasizes the irregular nature of dwarf galaxies.The H I and CO velocities agree well,so do the Hαvelocities.In total,we cataloged14molecular clumps in NGC4214. Our results from a virial mass analysis are compatible with a Galactic CO-to-H2 conversion factor for NGC4214(lower than what is usually found in metal–poor dwarf galaxies).Subject headings:galaxies:dwarf—galaxies:ISM1.IntroductionPrevious studies have shown that molecular gas is hard to study in dwarf galaxies, even in those who are known to be rich in neutral atomic gas(H I),such as dwarf irregulars (dIrrs)and blue compact dwarfs(BCDs).As carbon monoxide(CO)is the second most common molecular species after molecular hydrogen(H2),and the one most easily observed in its cool,unshocked state,it is usually used as a tracer of molecular gas.Relatively few dwarf galaxies have been detected in CO,though many have been observed(e.g.,Taylor, Kobulnicky&Skillman1998,Barone et al.2000).Dwarf galaxies are set apart from from other H I-rich galaxies by their low metallicity:Recent work suggested that the CO-to-H2 (‘X’)conversion factor is dependent upon the metallicity of the gas,with the conversion factor increasing as metallicity decreases(Arimoto,Sofue&Tsojimoto1996,Wilson1995, Verter&Hodge1995,Rubio et al.1991).Maloney&Black(1988)argued on theoretical grounds that at low metal abundances,the column density of CO will be lower,and there will be less self shielding from dissociating radiation.This will cause the size of the CO emitting region to shrink,while leaving the H2unaffected.Norman&Spaans(1997)predict that metallicities starting somewhere between0.03to0.1of the solar value are necessary before CO will be detectable.Observationally,Taylor et al.(1998)have found no dwarf galaxies with detectable CO emission at metallicities below∼0.15solar.However,at present there are too few dwarfs with both metallicity measurements and CO observations to determine reliably if there is indeed a cutoff,and where exactly it might take place.For the few dwarfs which have been detected in CO,following up those single dish detections with high spatial resolution interferometer data has proven enlightening.In some cases,the high resolution of interferometers allow the direct estimate of the total mass of resolved giant molecular clouds in nearby dwarf galaxies independent of the CO intensity, and from this afirst determination of the CO–to–H2conversion factor(e.g.,Wilson1995).A list of dwarf galaxies which have been observed with interferometers to date is compiled in Tab.1.E.g.,Taylor et al.(1999)used the Plateau de Bure interferometer to observe CO(1–0)and(2–1)in the post-starburst dwarf galaxy NGC1569.They found that the conversion factor was three times higher than in two other dwarf galaxies of the same metallicity(NGC6822and IC10),suggesting that metallicity is not the only influence upon the conversion factor.Motivated by this and other previous studies,we have observed another starburst dwarf galaxy,NGC4214,at Owens Valley.EDITOR:PLACE TABLE1HERE.NGC4214is a nearby dwarf galaxy currently experiencing high levels of massive star formation.Table2summarizes some of the properties of bining optical, NIR and UV data,Huchra et al.(1983)concluded it went through a burst of star formation a few×107yr ago.NGC4214has two main regions of star formation–a large complex of H II regions displaying a shell morphology in the galaxy center,and a smaller southern complex formed mainly of compact knots(MacKenty et al.2000).From age determinations of the H II regions and their spatial distribution,they argue that some of the younger star formation has been triggered by the interaction of the older H II regions with the surrounding ISM.At a distance of4.1Mpc(Leitherer et al.1996),it is one of the closest starbursting dwarf galaxies,and thus an excellent target for a high resolution study of the molecular ISM.Several groups have made single dish CO observations of NGC4214 (Tacconi&Young1985,Thronson et al.1988,Ohta et al.1993,Becker et al.1995,Taylor et al.1998).While the existence of CO emission away from the center of NGC4214can be inferred from the single dish data of Ohta et al.(who,however,fail to detect the central emission peak)and Taylor et al.,no analysis of the distribution of the molecular emission and its relation to the major star forming regions of the galaxy has been presented to date. Thronson et al.(1988)have also measured the dust mass in a region about4′in size using160µm data.The H I has been studied by Allsopp(1979)and McIntyre(1998)whofind a large extended H I disk and a complex velocityfield in the center.Based on the disturbed inner velocityfield,Allsopp(1979)argued that NGC4214is interacting with NGC4190,a dIrr with a projected distance of35kpc.EDITOR:PLACE TABLE2HERE.In Sec.2we describe our CO observations obtained with OVRO.In Sec.3we discuss the properties of the CO emitting regions detected by the interferometer,compare them to high–resolution VLA H I imaging as well as optical ground–based and HST imaging of NGC4214.In Sec.4we will discuss the conversion factor,the star formation efficiency, implications on the star formation threshold as derived from our CO and H I data as well as the interplay of the star formation regions on the surrounding ISM in NGC4214.We summarize our results in Sec.5.2.OVRO Observations2.1.Observations and Data ReductionWe observed NGC4214in the CO(1–0)transition using the OVRO millimeter interferometer in mosaicing mode in C and L configuration.In total,50hours were spent on source.The observational details are listed in Table3.Data were recorded using two correlator setups resulting in a velocity resolution of5and1.3km s−1(after online Hanning smoothing)with a total bandwidth of320and80km s−1,respectively.The absoluteflux calibration was determined by observing3C273,3C345and(in some observing periods) Neptune for approximately20minutes during each observing run.These calibrators and an additional noise source were used to derive the complex bandpass corrections.Thequasar1156+295,which is close to NGC4214on the sky,was used as secondary amplitude and phase calibrator.During the program,itsflux decreased from∼4.0Jy(Sep1999)to ∼1.0Jy(June2000).The data for each array were edited and calibrated separately with the MMA and the MIRIAD packages.The uv–data were inspected and bad data points due to either interference or shadowing between telescopes were removed,after which the data were calibrated.We Fourier transformed our C and L array observations separately for each pointing to assess their quality;subsequently we combined all data to form a single dataset which was used for mapping.The mosaicing and mapping was performed using the task MOSSDI in MIRIAD,i.e.a joint deconvolution of the entire mosaic was done to obtain the final data cubes.All the results presented in this paper are based on this combined dataset.We calculated2data cubes,one with natural weighting,leading to a resolution of 6.4′′×5.7′′,and another one with robust weighting(setting the robust parameter to0, Briggs1995)with a resolution of4.8′′×4.2′′.Thefinal rms noise in our1.3km s−1wide channels is40mJy beam−1(100mK)and47mJy beam−1(210mK)for the naturally and robustly weighted cube,respectively.We optimized the cleaning process by defining3clean boxes around the prominent emission regions.A summary of the observations can be found in Tab.3.Channel maps of the data cube are presented as contours in Fig.5.EDITOR:PLACE TABLE3HERE.2.2.Data ReductionWe distinguished CO emission from the noise using the following procedure:The cleaned1.3km s−1data cube was smoothed to a spatial resolution of9′′×9′′,and a2σ(80mJy beam−1)blanking was applied to the output cube.The resulting cube was searched for CO emission by looking for surviving features which were continuous in at least3 consecutive channels.This gave a cube containing only the CO emission at a resolution of 9′′.This cube was then applied as a mask to the original resolution naturally weighted cube to yield thefinal blanked data cube.The mask was also applied to the robustly weighted cube,and moment maps were made from the data of both weightings(Fig.1).The locations and sizes(FWHM)of the OVRO primary beams comprising our mosaic are shown in Fig.7.While we cannot be certain to have detected all the CO emission in NGC4214,we have probably mapped all the strong emission.Ourfirst set of observations included only the center mosaic position,but in those data we still detected both the northwestern and southeastern regions,even though they lay beyond the40%power radius of the primary beam,therefore if there were any similar CO features just beyond the beam areas shown in Fig.7,we would have detected them.It would be interesting to estimate the degree of missingflux due to the lack of short spacings in our interferometric map.This might give some indication of the amount and distribution of diffuse gas in and between the complexes.Unfortunately,there are no single dish observations of the CO distribution in NGC4214that are suitable for such a determination.The single dish beam is either very large(50′′−55′′)with a pointing center offset from the peak position of any complex,thus partially containing several complexes (Taylor et al.1998,Thronson et al.1988,Tacconi&Young1985),or the observed region is not sufficiently extended and spatially disjunct(Ohta et al.1993),or the observations were done in a different transition(Becker et al.1995for the central complex).EDITOR:PLACE FIGURE1HERE.EDITOR:PLACE FIGURE2HERE.3.Results3.1.CO Emitting RegionsAs seen in the moment map presented in Fig.1,there are three regions where we have detected CO emission in NGC4214–one at the center of the galaxy,another∼650pc to the southeast of the center,and the third∼760pc to the northwest.The southeastern region(SE)has the highest peak in CO emission,and is also the most centrally concentrated.The central region is extended in the east-west direction,with low intensity CO emission that has a diffuse appearance except for the concentration at its western end. The northwestern region(NW)is intermediate between the central and the southeastern ones,both in the degree of central concentration and the intensity of the emission.Each of the regions can be divided into discrete units(clumps)by considering both spatial and velocity information given in the data cube.To do that,we inspected the high–resolution robust–weighted cubes by looking at‘movies’as well as position–velocity (pv)diagrams.The clump decomposition is difficult in the NW and SW regions where many smaller units are blended.Fig.2shows a pv diagram of the NW region as an example of how we defined the clumps(the ellipses plotted in thisfigure are only used to label the clumps).The properties of the identified CO emitting clumps in NGC4214are compiled in Table4.C1–C8belong to the NW complex,C9belongs to the central complex,where also some diffuse emission is present(D1)and C10–C14are the constituents of the SE emission feature in NGC4214.For each clump,we derived a total intensity map,measured the diameter defined by the half intensity line and the velocity width:Tab.4lists the central coordinates(column2 and3),the central velocity of the clump(column4),the velocity width(FWHM,column 5),the COflux(column6)as well as the diameter of the clump(FWHM,column7)alongthe cut deconvolved for our beamsize(average of the orthogonal directions).In many cases, only an upper limit for the cloud size could be determined(as given by the size of the robust beam).We also attempted to estimate molecular masses by employing a Galactic X–factor(column8,see Sec.4.1)as well as assuming virialization(column9,but see the discussion in Sec.4.1).C1–C14are likely not Giant Molecular Clouds(GMCs).This can be seen in Fig.3, where we show the size-linewidth relation for GMCs in M33from Wilson&Scoville(1990), along with GMCs from M33,M31(Vogel et al.1987,Wilson&Rudolph1993),the SMC (Rubio et al.1993),IC10(Wilson1995),NGC1569(Taylor et al.1999)and NGC6822 (Wilson1994),and our complexes.The three complexes resolved by the OVRO beam clearly occupy a different region in the plot than GMCs in Local Group galaxies.They have similar linewidths,but are larger in size.The unresolved clouds may be individual GMCs, but their diameters are too small to measure with the synthesized beam of our data.The clumps we see fall in an area in the size-linewidth space which is in between the Local Group Galaxies and CO complexes in M83(Rand,Lord&Higdon1999),also plotted in Fig.3.Rand et al.argue that their objects are intermediate in size,linewidth,and molecular mass between the largest GMCs known in the Milky Way,and Giant Molecular Associations(GMAs)seen in nearby spiral galaxies(e.g.Rand&Kulkarni1990).Our resolved objects have smaller diameters compared to the Rand et al.objects–they are probably not GMAs.Also,GMAs tend to have masses of a few×107M⊙,whereas for an assumed Galactic CO-H2conversion factor,our clumps have masses that are lower by almost two orders of magnitudes.EDITOR:PLACE FIGURE3HERE.EDITOR:PLACE TABLE4HERE.parison to the H IStarbursting dwarf galaxies tend to be H I rich(e.g.Thuan&Martin1981,Taylor et al. 1995),and NGC4214is no exception.Fig.4shows the CO contours from Fig.1superposed on the total intensity map of the H I emission(corrected for primary beam attenuation). The high–resolution H I data are combined B,C and D configuration observations from the VLA,with a spatial resolution of8′′×8′′(see also McIntyre1998).The area we have mapped in CO covers only a small region at the center of NGC4214.The peak column density of H I in the map is∼3×1021cm−2,and the overall H I distribution shows a large degree of structure:H I holes and shells,which are common in dwarf galaxies(e.g.Puche et al.1992,Walter&Brinks1999),and some indication for spiral structure extending beyond the main optical disk.There is even a hint of a bar in the H I,running roughly position angle45◦(N→W)through the center of the galaxy.It is possible that both the weak spiral features and the bar are consequences of an interaction with the nearby dwarf galaxy NGC4190.There are a few dwarf galaxies such as Holmberg II(Puche at al.1992) and DDO47(Walter&Brinks2001)which show similar spiral features in H I only.The underlying process causing transient spiral structure may thus be fairly common.A global comparison between the H I and CO emission reveals differences between the various CO emitting regions.The NW region of CO emission falls on top of a local maximum in the H I column density,while the SE region,which is brighter and more centrally concentrated,corresponds to a dip in the H I column density near two local peaks. The central CO region is extended enough to cover several local minima and maxima in the H I,though the shape of the H I is elongated east-west,similar to the CO emission.It should be noted here that the strongest molecular emission is not necessarily found where the H I column densities are highest.Thisfinding emphasizes the‘irregular’nature of dIrrs.EDITOR:PLACE FIGURE4HERE.The CO and H I channel maps are compared in Fig.5.This representation of the data emphasizes more strongly the location of the SE CO emission in a small H I dip, and the association of the NW CO emission region with a local H I maximum roughly coincident in velocity.There is no clear correlation in the channel maps between CO and H I for the central CO feature.Fig.6is a pv-diagram through the three brightest CO peaks and compares the CO and H I velocities and intensities in and near these peaks. For the SE(offset:–0.75′)and central peak the CO linewidths are much narrower than the corresponding H I linewidths.In contrast,in the the NW peak(at offset0.7′)the CO linewidth is larger,probably due to substructure in the complex,and approaches the velocity width of the H I to within a factor of∼1.5.EDITOR:PLACE FIGURE5HERE.EDITOR:PLACE FIGURE6HERE.parison to Stellar Component and Regions of Star FormationStar formation occurs at densities where the gas is primarily molecular,so it is of interest to compare the molecular gas tracer,CO,to tracers of the stellar component. Because the ISM is readily ionized by the intense UV radiation from young stars in H II regions,Hαis often used as an indicator of the presence of recent star formation.Fig.7 compares the CO emission to an Hαimage(courtesy C.Martin).The left panel shows lacy,filamentary structure outside the immediate vicinity of the star forming regions.These are similar to those seen in other dwarf galaxies(e.g.NGC1569;Hunter et al.1993,Devost et al.1997,Martin1998),and are presumably caused by the energetic by-products of star formation,supernovae and stellar winds.Two of the three CO complexes are associated directly with the star forming regions (Fig.7,right).The SE CO clump appears cospatial with a large star forming complex containing several H II regions.The peak of the CO emission here is almost directly on top of one of the H II regions,and the shape of the CO complex with its slight extension in the north-south direction is very similar to that of the H II region.The central CO emission is associated with the largest region of Hαemission.Unlike the SE complex,the central CO complex is diffuse instead of centrally concentrated,and the peak of the CO emission here is not directly on top of the Hαemission,but is shifted to the west of it.Lower level CO emission is cospatial with the Hα,however,tracing out the chain of brighter H II regions running east-west in this area.In clear contrast,only very faint Hαemission is associated with the NW CO complex.At this point it is interesting to compare the velocities of the molecular gas with the ones of the ionized component(as seen in Hαspectra).Hartmann et al.(1986) presented the kinematics of a few H II regions within NGC4214–four of them correspond to regions where we also detected CO emission(their regions:H25/29,close to our NW CO complex;H36and H48,towards the central complex;and H55in the SE complex). In all cases do the systemic CO and Hαvelocities agree well(this is consistent with the results by Maiz–Apellaniz et al.1999).Only in H25/29(the weakest of the four Hαregions)is Hα–emission also present at velocities where no CO–emission is present(around v hel=250–270km s−1).Fig.8shows the contours of the CO emission superposed on a multi-band optical image created by the Hubble Heritage Team.Here,the quite different optical morphologies of the regions where CO emission is found are evident more dramatically.This composite image was created from different WFPC2pointings in the F336W(U),F502N([O III]), F555W(V),F656N(Hα),F702N(R),and F814W(I)filters.A detailed discussion of thesedata is published by MacKenty et al.(2000).Both Beck,Turner&Kovo(2000)and MacKenty et al.(2000)present high resolution VLA radio continuum images.There is6and20cm emission associated with the central and SE star forming complexes,but nothing is detected at the location of the NW CO emission(Turner2000,private comm.).This lack of radio continuum emission indicates that indeed very little star formation is associated with the NW molecular region(C1–C8) and that the presence of only very weak optical emission lines is not due to possible extinction.EDITOR:PLACE FIGURE7HERE.EDITOR:PLACE FIGURE8HERE.4.Discussion4.1.The CO-H2Conversion FactorTo determine the conversion factor with a reasonable degree of confidence,a discrete unit of emission,such as a GMC,must be distinguishable in the data,and there must be a method to determine the total mass associated with this feature which is independent of CO intensity.GMCs are often assumed to be gravitationally bound and in virial equilibrium, allowing the use of the virial theorem to determine the total mass.As discussed above, our spatial resolution will not resolve GMC–sized objects at the distance of NGC4214. For the sake of comparison,we decided to derive virial masses for our clumps anyway–in the discussion that follows we will keep in mind that at least some of them are likely notgravitationally bound objects.In the cases where the complex is not spatially resolved we took the beamsize(D∼90pc)as an upper limit for the actual diameter–then,the virial masses are probably upper limits(but see the discussion below).If the virial mass could be trusted,dividing it by the molecular mass derived using a Galactic CO-H2conversion factor would yield the actual conversion factor for each feature,in units of the Galactic value.The Galactic conversion factor is taken to be2.3×1020cm−2(K km s−1)−1(Strong et al.1988),with an additional explicit correction for helium(see caption in Table4).If the recalibrated conversion factor of1.6×1020cm−2(K km s−1)−1(Hunter et al.1997)or the new value advocated by Dame et al.2001(1.8×1020cm−2(K km s−1)−1)is used,all CO based masses are lower by∼30%.To facilitate comparison to previous work,the numbers we give refer to the’standard’value of Strong et al.The average conversion factor wefind for the few resolved features is compatible with the standard Galactic conversion factor(Tab.4)given all the uncertainties in deriving these numbers.The mean conversion factor is∼1.2±0.7×X MW,whether or not only the resolved clouds or all clouds are used for the estimate.This is a lower value than the conversion factor estimated for other dwarf galaxies with similar metallicities in the literature(2.7X MW for the dIrr IC10,<2.2X MW for the dIrr NGC6822,both in the Local Group,Wilson1995,and6.6X MW in the post-starburst dwarf galaxy NGC1569,Taylor et al.1999).This result is unexpected atfirst glance since one would expect the X–factor to be higher in this low metallicity galaxy(as suggested by other studies of low–metallicity dwarfs, see above).We can only speculate here why the results in NGC4214are different.We should like to note that the virial masses are strongly affected and possibly even dominated by the resolution(both spatially and kinematically)employed in the observations.E.g.,if we use our naturally weighted,5km s−1channel width data(i.e.,if we do not resolve thesubstructure)we derive a virial mass for the NW complex which is several times higher than the mass derived using X MW.This is due to fact that the dispersion in the NW region is rather large(∆v∼20km s−1)resulting in a much higher X–factor(because of the∆v2 dependence of the virial mass)–in other words our result would then be consistent with the values found in other low–metallicity galaxies!It should be also stressed here that deriving cloud properties(and hence virial masses)is not at all an objective process:e.g.,different radii have been used by different authors(e.g.,diameter defined by the half–intensity line;or by the contour containing90%of theflux)and the cloud sizes are not alwaysde–convolved for the beamsize.Dividing clumps into sub–clumps is also always subject to to some personal bias.It should be also noted at this point that convincing arguments have been presented (e.g.Bolatto et al.1999,Madden et al.1997)for a cloud structure in dwarf galaxies that significantly differs from what is found in metal-rich giant spirals.In this scenario,small CO cores are embedded in huge molecular envelopes which are devoid of CO.The molecular hydrogen in the envelopes coexists with C I and C II.E.g.,Madden et al.argue that in the dwarf irregular IC10the total molecular mass may be up to100times the mass contained in the CO cores.In this case our CO observations would only trace the cores of the clouds and the derived‘local’X–factor would be only valid for the central core of a molecular cloud,where it might indeed be close to the Galactic value,since the conditions in the core region are expected to be similar to those in a Galactic GMC.Unfortunately,based on the data presented here all this remains subject to speculation. In the case of NGC4214the matter might be investigated further by a comprehensive study of the CO,C I,C II,radio continuum and FIR emission,along the lines of the study Madden et al.performed for the(much closer)galaxy IC10.Based on our data,we will adopt X MW as the conversion factor for NGC4214in thefollowing,primarily for convenience.We keep in mind that this number is unfortunately only poorly determined and that our molecular gas mass estimated might be offby a factor of2–3.4.2.Star Formation efficiencyGiven the fact that the SE molecular cloud seems not yet to be disrupted by the ongoing burst of star formation(see the discussion in Sec.4.4),we can try to calculate the star formation efficiency for that region.From Table3of MacKenty et al.(2000)we derive a SFR for this region of0.066M⊙year−1(their region II,corrected for‘medium internal extinction’).For morphological reasons(e.g.the compact andfilled appearance of the H II regions,the lack of large cavities and the positional coincidence of exciting star clusters and Hαemitting knots),star formation presumably only recently started here–we adopted an age of∼3Million years(MacKenty et al.2000).This results in a total stellar mass of ∼2.0×105M⊙produced to date(assuming a constant SF rate).From our Table4,we derive a molecular mass of24×105M⊙(summing up regions C10–C14and adopting a Galactic X–factor,see the discussion in Sec.4.1).The star formation efficiency for the SE region is thus around8%,somewhat higher than what is usually adopted(see also,e.g., Wilson&Matthews1995and the discussion in Taylor et al.1999).4.3.Star formation ThresholdStudies of star forming dwarf galaxies suggest that the gas column density plays an important role in regulating star formation(Skillman et al.1987,Taylor et al.1994,Walter et al.1997,van Zee,Skillman&Salzer1998,Walter&Brinks1999).A threshold column density for the onset of star formation is thought to exist,at∼1021cm−2,above whichmassive star formation can proceed and below which it is suppressed.The physical reason for this threshold may be related to the density needed for the shielding of molecular gas from dissociating radiation(e.g.Skillman1987),or perhaps to disk instabilities(e.g. Kennicutt1989).For this kind of analysis,generally some measure of the spatial extent of the star formation activity(such as an Hαimage)is compared to the azimuthally averaged gas surface density(usually based on H I observations),or to the contours of an H I column density map.Because CO is difficult to detect in most dwarf galaxies,the previous studies have tended to ignore the presence of molecular gas entirely,or else to attempt to make a statistical correction by assuming a global H2to H I ratio(Hunter&Plummer1996). However,we have high resolution CO and H I data on NGC4214covering the star forming regions of the galaxy,so we can treat both the atomic and the molecular gas in the threshold analysis.We created a total gas column density map by combining the H I and CO moment maps,and using a Galactic CO-H2conversion factor(see Sec.4.1).Fig.9shows the contours for a column density of1,2and4×1021cm−2(corrected for an inclination of 30◦;McIntyre1997)for the total gas map and the H I map alone,superposed on the Hαimage from Fig.7.There is hardly any change in the lowest contour,indicating that at densities of1021cm−2the gas is primarily in atomic form.The higher contours,however, show marked changes from H I alone to total gas density,particularly at the location of the SE CO complex.In the total gas map,we see a peak in gas density corresponding to the group of H II regions,whereas if we consider the H I alone,the Hαemission sits in a local minimum of density,adjacent to two local peaks.The peak gas density in the total gas map at this complex is6.6×1021cm−2.In the central complex the addition of the molecular component also makes a difference,connecting two H I peaks that straddle the center of the star forming complex.。
