Study on continuous network design problem using simulated annealing and genetic algorithm
This article appeared in a journal published by Elsevier.The attached copy is furnished to the author for internal non-commercial research and education use,including for instruction at the authors institution
and sharing with colleagues.
Other uses,including reproduction and distribution,or selling or licensing copies,or posting to personal,institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article(e.g.in Word or Tex form)to their personal website or
institutional repository.Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
https://www.wendangku.net/doc/5a16240681.html,/copyright
Study on continuous network design problem using simulated
annealing and genetic algorithm
Tianze Xu a,*,Heng Wei b ,Zhuan-De Wang c
a
Department of Civil and Environmental Engineering,735Engineering Research Center,P.O.Box 210071,The University of Cincinnati,
Cincinnati,OH 45221-0071,USA
b
Department of Civil and Environmental Engineering,792Rhodes Hall,P.O.Box 210071,The University of Cincinnati,Cincinnati,OH 45221-0071,USA
c
School of Applied Mathematics,University of Electronic Science and Technology of China,Chengdu,Sichuan,610054,PR China
Abstract
In general,a continuous network design problem (CNDP)is formulated as a bi-level program.The objective function at the upper level is de?ned as the total travel time on the network,plus total investment costs of link capacity expansions.The lower level problem is formulated as a certain tra?c assignment model.It is well known that such bi-level program is non-convex and non-di?erentiable and algorithms for ?nding global optimal solutions are preferable to be used in solving it.Simulated annealing (SA)and genetic algorithm (GA)are two global methods and can then be used to determine the optimal solution of CNDP.Since application of SA and GA on continuous network design on real transportation network requires solving tra?c assignment model many times at each iteration of the algorithm,computation time needed is tremendous.It is important to compare the e?cacy of the two methods and choose the more e?cient one as reference method in practice.In this paper,the continuous network design problem has been studied using SA and GA on a simulated network.The lower level program is formulated as user equilibrium tra?c assignment model and Frank–Wolf method is used to solve it.It is found that when demand is large,SA is more e?cient than GA in solving CNDP,and much more computational e?ort is needed for GA to achieve the same optimal solution as SA.However,when demand is light,GA can reach a more optimal solu-tion at the expense of more computation time.It is also found that increasing the iteration number at each temperature in SA does not necessarily improve solution.The ?nding in this example is di?erent from [Karoonsoontawong,A.,&Waller,S.T.(2006).Dynamic continuous network design problem –Linear bilevel programming and metaheuristic https://www.wendangku.net/doc/5a16240681.html,work Modeling 2006Transportation Research Record (1964)(pp.104–117)].The reason might be the bi-level model in this example is nonlinear while the bi-level model in their study is linear.
ó2008Published by Elsevier Ltd.
Keywords:Continuous network design;Tra?c assignment;User equilibrium;Simulated annealing;Genetic algorithm
1.Introduction
The network design problem is to improve transporta-tion network by selecting facilities (for example,entire-lane or new link)to add to a transportation network,or to deter-mine capacity enhancements of existing facilities of a trans-portation network.Continuous network design problem (CNDP)is concerned with divisible capacity enhancements
(Friesz,Cho,Mehta,Tobin,&Anandalingam,1992),for example,altering lane width,median and shoulder area.In general,a continuous network design problem (CNDP)is formulated as a bi-level program.The upper level could be a multi-objective model or a model with objective func-tion de?ned as the sum of total travel time on the network and total investment costs of link capacity expansions.The lower level problem is formulated as a certain tra?c assign-ment model,which can be either a static tra?c assignment model or dynamic tra?c assignment model.
Determining the global optimal solution is of great importance in CNDP.It is well known that such a bi-level
0957-4174/$-see front matter ó2008Published by Elsevier Ltd.doi:10.1016/j.eswa.2008.01.071
*
Corresponding author.Tel.:+1(513)7513128;fax:+1(513)5562599.E-mail address:xut@https://www.wendangku.net/doc/5a16240681.html, (T.Xu).
https://www.wendangku.net/doc/5a16240681.html,/locate/eswa
Available online at https://www.wendangku.net/doc/5a16240681.html,
Expert Systems with Applications 36(2009)
2735–2741
Expert Systems with Applications
model is non-convex and non-di?erentiable.Global methods such as simulated annealing(SA)and genetic algorithm(GA)are preferable to be used in solving the model.Both SA and GA require evaluating the objective function value of the upper level model many times at each iteration of the algorithm.The lower level tra?c assign-ment model has to be solved in order to evaluate the upper level objective function value.The computation time for solving a tra?c assignment model of a real transportation network is considerably large.Thus the application of SA and GA to solve continuous network design requires tre-mendous computation time.It is important to compare the e?cacy of the two methods and choose the more e?-cient one as reference method in practice.
In this paper,the continuous network design problem has been studied using SA and GA on a simulated network. The lower level program is formulated as static user equi-librium tra?c assignment model and Frank–Wolf method is used to solve it.The computation time needed of each method for?nding a optimal solution is compared.
The paper is organized as follows:Section2reviews lit-erature and introduces previous study on CNDP.Section3 presents problem formulation of CNDP.Section4intro-duces GA.Section5introduces SA.Section6presents a numerical example and compares the e?ciency of GA and SA.Section7concludes the paper.
2.Literature review
CNDP has been formulated as bi-level model of di?er-ent kind.The upper level is formulated either as a multi-objective model(Fan&Machemehl,2006;Friesz et al., 1993;etc.)or a model with objective function de?ned as the sum of total travel time on the network and total investment costs of link capacity expansions(Chiou, 2005;Friesz et al.,1992;Karoonsoontawong&Waller, 2006;etc.).The lower level problem is formulated as a cer-tain tra?c assignment model,such as static user equilib-rium(Ban,Liu,Lu,&Ferris,2006;Chiou,2005;Friesz et al.,1992),variable demand equilibrium(Chen&Chou, 2006),stochastic user equilibrium(Chen,Subprasom,& Ji,2006;Davis,1994),user-optimal dynamic tra?c assign-ment(UO DTA)model(Karoonsoontawong&Waller, 2006),etc.
Di?erent methods have been used to solve CNDP mod-els.Study by Friesz(1992)shows that simulated annealing (SA)is superior to iterative optimization-assignment algo-rithm(IOA),Hooke–Jeeves algorithm(HJ),equilibrium decomposed optimization(EDO),and modular in-core nonlinear system(MINOS)in?nding global optimal solu-tion when the lower model is formulated as static user equi-librium.Davis(1994)used the generalized reduced gradient method and sequential quadratic programming to?nd the optimal solution of CNDP.Chiou(2005)proposed four gradient-based methods to solve the CNDP.The four gra-dient-based methods are Gradient Projection method(GP), Conjugate Gradient projection method(CG),Quasi-NEW-ton projection method(QNEW),and PARATAN version of gradient projection method(PT).He applied the meth-ods to the sixteen link network used by Friesz(1992)and Sioux Falls city network and compared the four methods with other methods including IOA,HJ,EDO,MINOS, SA,sensitivity analysis based algorithm(SAB),and Aug-mented Lagrangian algorithm(AL).The comparison shows that CG and QNEW outperforms all other methods except SA.SA outperforms all other methods in all cases. But the di?erence between CG,QNEW and SA is not large.Ban,Liu,Ferris,and Ran(2006)presented a Relax-ation method(RELAX)to solve CNDP when the lower level is a nonlinear complementary problem.They also applied the methods to the same network as Chiou’s study and compared RELAX with other methods including IOA, EDO,MINOS,SA,and AL.Their study shows SA and RELAX outperforms all other methods.In total three cases,SA outperforms RELAX once and RELAX out-performs SA twice.But the di?erence is very small. Karoonsoontawong and Waller(2006)used simulated annealing(SA),genetic algorithm(GA),and random search(RS)to?nd optimal solutions of CNDP when it is modeled as a linear bi-level programming with a lower level of dynamic user-optimal tra?c assignment model.His study shows that GA outperforms the others in the test problems.
The objective of this paper is to study continuous net-work design problem using SA and GA when the lower level is modeled as static user equilibrium(UE).Speci?-cally,the e?ciency of the two global methods will be com-pared.Though Karoonsoontawong and Waller(2006)’s study has showed that GA outperforms SA when the lower level is dynamic tra?c assignment,it is still necessary to compare the e?cacy of SA and GA when the lower level is modeled as static tra?c assignment because the bi-level static model is non-convex nonlinear and is di?erent from linear bi-level programming in their study.Since the time needed for solving the lower tra?c assignment model accounts for major proportion of computation time of the whole problem,the UE assignment number will be used to measure the computation cost.It is more reasonable to use UE assignment number instead of CPU time to mea-sure computation cost of either method since it is indepen-dent of network and is more comparable.
3.Problem formulation
The continuous network design problem(CNDP)with ?xed demand static user equilibrium?ow constraint is for-mulated as:
min Tex;yT?
X
a2A
et aex aeyT;y aTx aeyTtq g aey aTTe1T
s:t:06y
a
6u
a
8a2A
where x(y)is the equilibrium?ow de?ned by the following ?xed demand static user equilibrium problem:
2736T.Xu et al.,/Expert Systems with Applications36(2009)2735–2741
min zexT?
X
a Z x a
t aex;y aTd xe2T
s:t:X
l
f rs
l
?q rs8r2R;s2S
x a?
X
rs
X
k2K rs f rs
k
d rs
a;k
8r2R;s2S;a2A;k2K rs
f rs
k
P08r2R;s2S;k2K rs
and where
A the set of links in the network
R the set of origins
S the set of destinations
D the vector of?xed OD pair demands,D=[D rs]
"r2R,s2S
K rs the set of paths between OD pair rs"r2R,s2S
f the vector of path?ows,f??f rs
k "r2R,s2S,
k2K rs
x the vector of equilibrium link?ows,x=[x a] "a2A
y the vector of link capacity expansions,y=[y a] "a2A
u the vector of upper bound for link capacity expan-sions,u=[u a]"a2A
q the conversion factor from investment cost to tra-vel times
t the vector of link travel times,t=[t a(x a,y a)] "a2A
g the vector of investment costs,g=[g a(y a)]"a2A
D the link-path incidence matrix,D??d rs
a;k ,where
d rs a;k ?1if link a is on th
e k th route connecting ori-
gin r and destination s,and d rs
a;k
?0otherwise
In this bi-level model,the upper level is non-convex and non-di?erentiable in y which is de?ned by the lower level model.Below we will introduce two global methods includ-ing SA and GA to solve the bi-level model.The e?cacy of the two methods will be compared.
4.Genetic algorithm(GA)
A genetic algorithm(GA)is a global search heuristic technique used in computing to?nd true or approximate solutions to optimization problems.It uses techniques such as inheritance,mutation,selection,and crossover which are inspired by evolutionary biology.Genetic algorithms are implemented as a computer simulation in which a popula-tion of chromosomes of candidate solutions to an optimi-zation problem evolves toward better solutions.Solutions can be represented in binary or real-coded.The evolution usually starts from a population of randomly generated individuals and happens in generations.In each generation, the?tness value of every individual in the population is evaluated,multiple individuals are randomly selected from the current population(based on their?tness value)and modi?ed(recombined and possibly randomly mutated)to form a new population.The new population is then used in the next iteration of the https://www.wendangku.net/doc/5a16240681.html,monly,the algo-rithm terminates when either a maximum number of gener-ations has been produced,or a satisfactory?tness level has been reached for the population.
A real-coded genetic algorithm with multiple crossovers was used in this paper.The algorithm is well described by Chang(2007).A brief summary of the algorithm is given below for completeness.
4.1.Initialization
A search space should be?rst de?ned.All genes in the chromosome will be operated and evaluated in this con-strained space.Let population size N represent the number of original chromosomes.The initial solution was gener-ated randomly in the feasible region.Once a generated chromosome by genetic operations goes beyond the bound, the original chromosome will be retained.
4.2.Reproduction
A simpler tournament selection is adopted to decide whether a chromosome can reproduce or not based on its ?tness value(z).The tournament selection says that p r N chromosomes with minimum z values are more added into the population and p r N chromosomes with maximum z val-ues are discarded from the population,where p r is the prob-ability of reproduction.The resulting population has the same size with the original one.After the selection,all chromosomes are completely put in the mating pool.
4.3.Crossover
The N chromosomes are randomly divided into N/2 pairs.Assume that both y and y are selected and c is a ran-dom number chosen from[0,1].If c>p c,where p c is prob-ability of crossover,then the following crossover operations for y and y are performed
if zeyT y0? ytreyà yT; else y0?ytre yàyT; y0? ytre yàyT; where z(y)and ze yTare?tness values of chromosomes y and y,respectively,y0and y0are the resulted children chro-mosomes,and r2[0,1]is a random number determining the crossover grade of these two.If c 4.4.Mutation The mutation operation follows the multiple crossover and provides a possible mutation on some selected chro-mosomes.Only p m N random chromosomes in the current population are chosen to be mutated.The formula of mutation operation for a selected y is given by T.Xu et al.,/Expert Systems with Applications36(2009)2735–27412737 y0?ytsáh;e4Twhere s is a small positive constant and h is a random per-turbation vector to produce small disturbances on y. The overall real-coded GA on CNDP is summarized as follows: Step1:Create a population with N chromosomes,which are randomly generated from the feasible region. Step2:For each chromosome in the population,solve program(2)and evaluate the?tness value with(1). Step3:If the prespeci?ed number of generations is reached or there is a chromosome in population with?tness value less than prespeci?ed error,stop; else,continue. Step4:Reproduction. Step5:Perform multiple crossover based on(3). Step6:Perform mutation based on(4).(If the resulting chromosome during these operations is outside the region,the original one is retained). Step7:Go back to Step2. 5.Simulated annealing(SA) Simulated annealing(SA)is a stochastic gradient method for the global optimization problem.It was origi-nally developed to simulate the annealing process.It starts from an initial solution at a high temperature,and makes a series of moves according to annealing schedule.The change in the objective function values(D E)is computed at each move.If the new solution results in decreased objective function value,it is accepted with probability1. If the new solution yields increased objective function value,it is accepted with1a small probability p,which is de?ned as P(D E)=exp(àD E/k B T),where k B is Boltz-mann’s constant and T is the current temperature.By accepting worse solutions with a certain probability,SA can avoid being trapped on a local optimum.SA repeats this process M times at each temperature to reach the ther-mal equilibrium,where M is a control parameter,also known as Markov length(Wu,Chang,&Chung,2007). The parameter T is gradually decreased as SA proceeds until the stopping condition is met.It terminates,when either the optimal solution is attained or the problem becomes frozen at a local optimum that cannot be improved. The algorithm is well described by Liu(2001).A brief summary of the algorithm is given below for completeness. Some factors need to be considered when designing the SA algorithm are introduced?rst. 5.1.Initialization An initial solution is generated randomly from the feasi-ble region.An initial temperature should be high enough to allow all candidate solutions to be accepted.5.2.Markov length The iteration number M used in each temperature.This number should be set appropriately high for the objective function values to reach Boltzmann distribution. 5.3.Cooling schedule Cooling schedule is the rate at which the temperature is reduced.In this paper,0.8is used at the?rst12tempera-ture reductions.0.8means the temperature of the next stage is0.8times the current temperature.0.5is used after the12th temperature reduction. 5.4.Step size Step size at each move should be decreased with the reduction of temperature.The feasible solutions at lower temperature are close to optimal solution.When tempera-ture is low,the stochastic search tends to be deterministic search.So if step size is too large,at low temperature,some feasible solutions will be rejected,thus computation time will be wasted. 5.5.Neighboring solutions Neighboring solutions are the set of feasible solutions that can be generated from the current solution.Each fea-sible solution can be directly reached from current solution by a move and resulted neighboring solution. 5.6.Stopping criteria The algorithm stops when the number of temperature transitions reaches a prespeci?ed number,or when the tem-perature is reduced to a threshold,or when the neighbor solution was not improved after a period. The algorithm of SA on CNDP is summarized as: Step1:Initialization. 1.1Generate an initial feasible solution y0,solve program (2)and evaluate the objective function value z(y0) with(1)based on y0,let y=y0. 1.2Set Markov length M,initial step size a0,initial tem- perature T0,error e,let a?a0;T?T0: 1.3Set the outer iteration counter n=1. Step2:For the given T,perform the following: 2.1Set the inner iteration counter k=1. 2.2Let^y?yta U;n?nt1,where U=(...,U j, (i) a random vector and U j is independently uniformly distributed on[à ??? 3 p , ??? 3 p ]. 2.3Solve program(2)and evaluate the objective function value ze^yTwith(1)based on^y. 2738T.Xu et al.,/Expert Systems with Applications36(2009)2735–2741 2.4Set D z ?z e^y Tàz ey T,if D z <0,y ?^y ;else let y ?^y with probability P (D z )=exp (àD z /T ). 2.5k =M ?If k =M ,go to step 3;else k =k +1,return to 2.2.Step 3:Stop or not 3.1T T =0.8T ,n =n +1,return to Step 2. 6.A numerical example In this section,the e?cacy of the GA and SA for contin-uous network design programs of the simulated network is described and compared.The network is shown in Fig.1.It consists of 18links and 6nodes.This network is a modi?ed network from the network used by Friesz (1992).The tra-vel demand for this network includes three travel demand cases and is shown in Table 1.Case 1is for light demand.Case 2and case 3are for heavy demand.The parameters for the network are shown in Table 2.The travel time func-tion of each link is de?ned as t a (x a ,y a )=A a +B a (x a /(K a +y a )4).The parameters used in SA are summarized in Table 3.The parameters used in GA are summarized in Table 4. The comparison of SA and GA are shown in Table 5.In case 1in which demand is light,SA reaches an optimal value of 205.8907and 15,000UE assignments need to be made to reach this optimal;GA reaches an optimal value of 191.2557and 50,000UE assignments need to be made to reach this optimal.In case 2and case 3,SA reach opti-mal value of 505.3919and 739.5390and the corresponding UE assignment number is 42,500and 22,500;GA reach optimal value of 515.0984and 744.3962and 50,000UE assignments are needed. The comparison is made clearer by plotting the optimal objective function value z *versus UE assignment number (UE#)needed to reach the optimal for both SA and GA.The plot is shown in Figs.2–4,which are for case 1,case 2and case 3,respectively.Some observations can be made based on the plot.When demand is light as in case 1,GA can reach a more optimal value than SA at the expense of more UE assignments.When demand is heavy as in case 2and case 3,the optimal values found by SA and GA are about the same,but the UE assignment number for GA is much more than that in SA.In three cases,the objective function value by GA decreases consistently with the increase of UE assignment number;however, solutions Fig.1.Eighteen link network. Table 1 Travel demand for the example Case Travel demand from 1to 6q 16Travel demand from 6to 1q 61Total ?ow 15101521020303 15 25 40 Table 2 Network parameters for the example Arc a A a B a K a d a 11103222510333395442044555039622021711014811103928452103335119226124106813425445142332031555161661 4.61175945218 5 9 45 2 Table 3 Parameters for simulated annealing algorithm in the example Parameter Parameter value Initial temperature 500Final temperature 0.0052Temperature reduction number 25 Iterations at each temperature M 300,400,500,...,1800,1900,2000 Step at each iteration a n ?n à1 n a n à1Upper bound on y a 20 Table 4 Parameters for genetic algorithm in the example Parameter Parameter value Population size N 10Probability of reproduction p r 0.2Probability of crossover p c 0.2Probability of mutation p m 0.2Step in mutation operation s 0.5Upper bound on y a 20 Number of generations 750,1000,1250,...,4500,4750,5000 T.Xu et al.,/Expert Systems with Applications 36(2009)2735–27412739 by SA do not show obvious improvement with the increase of UE assignment number(which is made by increasing Markov length or the iteration number at each tempera-ture).In SA,solution found with lower Markov length (for example300)is as optimal as solutions found with higher Markov length(for example1800).This shows that increasing the iteration number at each temperature in SA does not necessarily improve solution and an appropriately low Markov length should be used to save computation time.It is interesting to note the performance of SA and GA in this example is di?erent from the result in Karo-onsoontawong and Waller’s study(2006)because the bi-level model in this example is nonlinear while the bi-level model in their study is linear.7.Conclusion In this paper,the continuous network design problem (CNDP)has been studied using SA and GA on a simulated network.The CNDP is modeled as a bi-level non-convex nonlinear program.The objective function at the upper level is de?ned as the total travel time on the network,plus total investment costs of link capacity expansions.The lower level program is formulated as user equilibrium traf-?c assignment model.Two global methods including SA and GA are used to?nd the optimal solution of the CNDP program and the e?cacy of the two methods are compared. It is found that when demand is light,SA is more e?cient than GA in solving CNDP and much more computational e?ort is needed for GA to?nd the same optimal solution as SA.However,when demand is light,GA can reach a more optimal solution at the expense of more computation time. It is also found that increasing the iteration number at each temperature in SA does not necessarily improve solution. Our?nding is di?erent from Karoonsoontawong and Wal- Table5 Comparison of SA and GA Case1Case2Case3 SA GA SA GA SA GA y1000000 y20.46880 1.7313 2.20479.124411.9861 y30.6543011.770010.613318.122016.2407 y4000000 y5000000 y6 6.5331 4.4698 4.7528 6.6830 4.9850 5.4073 y70.796500.141500.11270 y80.246000.78230.0026 1.5768 6.0434 y9000000 y10000000 y11000000 y12000000 y13000000 y140.84210 5.9448 1.216711.664912.2779 y150.13610 1.5119 6.3021 2.97160.8197 y167.34477.540318.449611.928919.719119.9897 y17000000 y18000000 z*205.8907191.2557505.3919515.0984739.5390744.3962 UE#15,00050,00042,50050,00022,50050,000 2740T.Xu et al.,/Expert Systems with Applications36(2009)2735–2741 ler’s study(2006).The reason might be the bi-level model in this example is nonlinear while the bi-level model in their study is linear. References Ban,X.G.,Liu,H.X.,Lu,J.G.,&Ferris,M.C.,2006.Decomposition scheme for continuous network design problem with asymmetric user https://www.wendangku.net/doc/5a16240681.html,work Modeling2006Transportation Research Record (1964)(pp.185–192). Ban,J.X.,Liu,H.X.,Ferris,M.C.,&Ran,B.(2006).A general MPCC model and its solution algorithm for continuous network design problem.Mathematical and Computer Modelling,43(5–6),493–505. Chang,W.-D.(2007).A multi-crossover genetic approach to multivariable PID controllers tuning.Expert Systems with Applications,33,620–626. Chen,H.K.,&Chou,H.W.(2006).Reverse supply chain network design https://www.wendangku.net/doc/5a16240681.html,work Modeling2006Transportation Research Record (1964)(pp.42–49). Chen,A.,Subprasom,K.,&Ji,Z.W.(2006).A simulation-based multi-objective genetic algorithm(SMOGA)procedure for BOT network design problem.Optimization and Engineering,7(3),225–247. Chiou,S.W.(2005).Bilevel programming for the continuous transport network design problem.Transportation Research Part B–Method-ological,39(4),361–383.Davis,G.A.(1994).Exact local solution of the continuous network design problem via stochastic user equilibrium assignment.Transportation Research Part B–Methodological,28(1),61–75. Fan,W.,&Machemehl,R.B.(2006).Optimal transit route network design problem with variable transit demand:Genetic algorithm approach.Journal of Transportation Engineering-ASCE,132(1),40–51. Friesz,T.L.,Anandalingam,G.,Mehta,N.J.,Nam,K.,Shah,S.J.,& Tobin,R.L.(1993).The multiobjective equilibrium network design problem revisited–A simulated annealing approach.European Journal of Operational Research,65(1),44–57. Friesz,T.L.,Cho,H.J.,Mehta,N.J.,Tobin,R.L.,&Anandalingam,G. (1992).A simulated annealing approach to the network design problem with variational inequality constraints.Transportation Sci-ence,18–26. Karoonsoontawong,A.,&Waller,S.T.(2006).Dynamic continuous network design problem–Linear bilevel programming and metaheu-ristic https://www.wendangku.net/doc/5a16240681.html,work Modeling2006Transportation Research Record(1964)(pp.104–117). Liu,C.(2001).Advanced tra?c planning.People’s Transportation Press (pp.345–346). Wu,T.-H.,Chang,C.-C.,&Chung,S.-H.(2007).A simulated annealing algorithm for manufacturing cell formation problems.Expert Systems with Applications. T.Xu et al.,/Expert Systems with Applications36(2009)2735–27412741 《葛底斯堡演讲》三个中文译本的对比分析 葛底斯堡演讲是林肯于19世纪发表的一次演讲,该演讲总长度约3分钟。然而该演讲结构严谨,富有浓郁的感染力和号召力,即便历经两个世纪仍为人们津津乐道,成为美国历史上最有传奇色彩和最富有影响力的演讲之一。本文通过对《葛底斯堡演讲》的三个译本进行比较分析,从而更进一步加深对该演讲的理解。 标签：葛底斯堡演讲,翻译对比分析 葛底斯堡演讲是美国历史上最为人们所熟知的演讲之一。1863年11月19日下午,林肯在葛底斯堡国家烈士公墓的落成仪式上发表献词。该公墓是用以掩埋并缅怀4个半月前在葛底斯堡战役中牺牲的烈士。 林肯是当天的第二位演讲者,经过废寝忘食地精心准备,该演讲语言庄严凝练,内容激昂奋进。在不足三分钟的演讲里,林肯通过引用了美国独立宣言中所倡导的人权平等,赋予了美国内战全新的内涵,内战并不仅是为了盟军而战,更是为了“自由的新生(anewbirthoffreedom)”而战,并号召人们不要让鲜血白流,要继续逝者未竞的事业。林肯的《葛底斯堡演讲》成功地征服了人们,历经多年仍被推崇为举世闻名的演说典范。 一、葛底斯堡演说的创作背景 1.葛底斯堡演说的创作背景 1863年7月1日葛底斯堡战役打响了。战火持续了三天,战况无比惨烈,16万多名士兵在该战役中失去了生命。这场战役后来成为了美国南北战争的一个转折点。而对于这个位于宾夕法尼亚州,人口仅2400人的葛底斯堡小镇,这场战争也带来了巨大的影响——战争遗留下来的士兵尸体多达7500具,战马的尸体几千具,在7月闷热潮湿的空气里,腐化在迅速的蔓延。 能让逝者尽快入土为安,成为该小镇几千户居民的当务之急。小镇本打算购买一片土地用以兴建公墓掩埋战死的士兵,然后再向家属索要丧葬费。然而当地一位富有的律师威尔斯(DavidWills)提出了反对意见,并立即写信给宾夕法尼亚州的州长,提议由他本人出资资助该公墓的兴建,该请求获得了批准。 威尔斯本打算在10月23日邀请当时哈佛大学的校长爱德华(EdwardEverett)来发表献词。爱德华是当时一名享有盛誉的著名演讲者。爱德华回信告知威尔斯,说他无法在那么短的时间之内准备好演讲,并要求延期。因此,威尔斯便将公墓落成仪式延期至该年的11月19日。 相比较威尔斯对爱德华的盛情邀请,林肯接到的邀请显然就怠慢很多了。首先,林肯是在公墓落成仪式前17天才收到邀请。根据十九世纪的标准,仅提前17天才邀请总统参加某一项活动是极其仓促的。而威尔斯的邀请信也充满了怠慢, 话说宝玉在林黛玉房中说"耗子精"，宝钗撞来，讽刺宝玉元宵不知"绿蜡"之典，三人正在房中互相讥刺取笑。 杨宪益：Pao-yu，as we saw, was in Tai-yu?s room telling her the story about the rat spirits when Pao-chai burst in and teased him for forgetting the “green wax” allusion on the night of the Feast of Lanterns. 霍克斯: We have shown how Bao-yu was in Dai-yu?s room telling her the story of the magic mice; how Bao-Chai burst in on them and twitted Bao-yu with his failure to remember the …green wax? allusion on the night of the Lantern Festival; and how the three of them sat teasing each other with good-humored banter. 对比分析：杨宪益和霍克斯在翻译“耗子精”采用来了不同的处理方法，前者使用了异化”rat spirits”，后者用的是归化法”magic mice”,使用归化法更受英美读者的亲乃。但是二者同时采用了增译法，增添了the story，原文并没有。在翻译“宝玉不知绿烛之典”的“不知”，英文1用的是“forgetting”，而译文2用的是“with failure to ”，显然译文2更符合英美的表达习惯。 那宝玉正恐黛玉饭后贪眠，一时存了食，或夜间走了困，皆非保养身体之法。幸而宝钗走来，大家谈笑，那林黛玉方不欲睡，自己才放了心。 杨宪益：Pao-yu felt relieved as they laughed and made fun of each other, for he had feared that sleeping after lunch might give Tai-yu indigestion or insomnia that night, and so injure her health. Luckily Pao-chai?s arrival and the lively conversation that followed it had woken Tai-yu up. 霍克斯: Bao-yu had been afraid that by sleeping after her meal Dai-yu would give herself indigestion or suffer from insomnia through being insufficiently tired when she went to bed at night, but Bao-chai?s arrival and the lively conversation that followed it banished all Dai-yu?s desire to sleep and enabled him to lay aside his anxiety on her behalf. 对比分析：译文一对原文语序进行了调整，先说了“放心”，再说“担心”，但并不如不调整顺序的逻辑强。译文二只是用了一个“but”就把原文意思分层了两层，逻辑更加清晰，符合西方人注重逻辑的习惯。原文中的“谈笑”是动词，而两个译文版本都是译的“the lively conversation”，是名词，体现了汉语重动态，英文重静态的特点。 忽听他房中嚷起来，大家侧耳听了一听，林黛玉先笑道："这是你妈妈和袭人叫嚷呢。那袭人也罢了，你妈妈再要认真排场她，可见老背晦了。" 杨宪益：Just then, a commotion broke out in Pao-yu?s apartments and three of th em pricked up their ears. “It?s your nanny scolding Hai-jen,” announced Tai-yu. “There?s nothing wrong with Hai-jen, yet your nanny is for ever nagging at her. Old age has befuddled her.” 《傲慢与偏见》（节选一） Pride and Prejudice by Jane Austen (An Except from Chapter One) 译文对比分析 节选文章背景：小乡绅贝内特有五个待字闺中的千金，贝内特太太整天操心着为女儿物色称心如意的丈夫。新来的邻居宾利(Bingley)是个有钱的单身汉，他立即成了贝内特太太追猎的目标。 1.It’s a truth u niversally acknowledged, that a single man in possession of a good fortune, must be in want of a wife. 译文一：凡是有钱的单身汉，总想娶位太太，这已经成了一条举世公认的道理。译文二：有钱的单身汉总要娶位太太，这是一条举世公认的真理。 2.However little known the feelings or views of such a man may be on his first entering a neighborhood, this truth is so well fixed in the minds of the surrounding families that he is considered as the rightful property of some one or other of their daughters. 译文一：这样的单身汉，每逢新搬到一个地方，四邻八舍虽然完全不了解他的性情如何，见解如何，可是，既然这样的一条真理早已在人们心中根深蒂固，因此人们总是把他看作是自己某一个女儿理所应得的一笔财产。 译文二：这条真理还真够深入人心的，每逢这样的单身汉新搬到一个地方，四邻八舍的人家尽管对他的性情见识一无所知，却把他视为某一个女儿的合法财产。 3.”My dear Mr. benne” said his lady to him one day ,”have you heard that nether field park is let at last?” 译文一：有一天班纳特太太对她的丈夫说：“我的好老爷，尼日斐花园终于租出去了，你听说过没有？” 译文二：“亲爱的贝特先生”一天，贝纳特太太对先生说：“你有没有听说内瑟费尔德庄园终于租出去了：” 4.Mr. Bennet replied that he had not. 译文一：纳特先生回答道，他没有听说过。 译文二：纳特先生回答道，没有听说过。 5.”But it is,” returned she:” for Mrs. long has just been here, and she t old me all about it.” 译文一：“的确租出去了，”她说，“朗格太太刚刚上这来过，她把这件事情的底细，一五一十地都告诉了我。” 译文二：“的确租出去了，”太太说道。“朗太太刚刚来过，她把这事一五一十地全告诉我了。” 《乡愁》两个英译本的对比分析 曹菊玲 （川外2013级翻译硕士教学1班文学翻译批评与鉴赏学期课程论文） 摘要：余光中先生的《乡愁》是一首广为人知的诗歌，曾被众多学者翻译成英文，本文将对赵俊华译本和杨钟琰译本进行对比分析，主要从原作的分析、译文的准确性、译文风格以及对原文意象的把握入手进行赏析，指出两译本的不足和精彩之处。 关键词：《乡愁》；原诗的分析；准确性；风格；意象的把握 1.原诗的分析 作者余光中出生于大陆，抗战爆发后在多地颠沛流离，最后22岁时移居台湾。20多年都未回到大陆，作者思乡情切，于上世纪70年代创作了这首脍炙人口的诗歌，但是当时大陆与台湾关系正处于紧张时期，思而不得的痛苦与惆怅包含于整首诗中。 全诗为四节诗，节与节之间完全对称，四句一节，共十六句，节奏感、韵律感很强，读起来朗朗上口。而且这首诗语言简洁，朴实无华，但是却表达了深刻的内涵。从整体上看，作者采取层层递进的方式，以时空的隔离和变化来推进情感的表达，最后一句对大陆的思念，一下子由个人哀愁扩大到国家分裂之愁，因而带有历史的厚重感。 【原诗】 乡愁 余光中 小时候／乡愁是一枚小小的邮票／我在这头／母亲在那头 长大后／乡愁是一张窄窄的船票／我在这头／新娘在那头 后来啊／乡愁是一方矮矮的坟墓／我在外头／母亲在里头 而现在／乡愁是一湾浅浅的海峡／我在这头／大陆在那头 【译诗】 (1) 赵俊华译 Homesick As a boy, I was homesick for a tiny stamp, —I was here, Mom lived alone over there. When grow up, I was homesick for a small ship ticket. —I was here, 《落花生》两英译本的对比分析 【摘要】本文对张培基先生和刘士聪先生对《落花生》的英译本从接受美学的角度进行了对比分析。通过分析原文的写作目的，风格和语言特点从而对翻译的分析打下良好基础。举例对翻译的段落进行了对比分析，反映了译者的期待视野影响翻译目的，职业经历和翻译的基本观点影响期待视野。这一点对于文本不确定性和语义空白点的具体化十分重要。 【关键词】《落花生》；张培基；刘士聪；接受美学；翻译对比 一、原文的介绍 《落花生》是中国著名作家许地山创作的一篇具有深远意义的散文。他回忆了童年时发生的一件小事。父亲通过一件关于花生的小事讲述了生活哲理。 原文具有以下几个功能： 1、信息功能：它描述了童年发生的一件事及花生的特点和用途。 2、审美功能：文章的语言简单朴素清新自然。语言特点，内容和风格相得益彰，形成了浓厚的艺术吸引力。 3、表达功能：这篇散文表达了作者崇高的思想，即便是在动荡浑浊的旧时代，还应保持个人节操。 4、祈使功能：虽然描写的是不起眼的花生，却在字里行间向大家传递了人生哲学。那就是“那么，人要做有用的人，不要做伟大、体面的人了。” 二、两英译本对比分析 张先生和刘先生在翻译时都保留了原文的风格和特点。但他们的翻译风格还是有些不同的。 他们的期待视野影响其翻译目的。他们的职业经历影响他们翻译期待视野。张先生和刘先生都曾在出版社做过编辑，都潜心研究翻译事业。但是不同的是刘先生还有翻译教学的经历。和张先生相比，他更算的上是一个翻译教育学者。他投身于翻译教学和其他英语相关的学科研究中。从以上可以推断出，可能刘先生在翻译过程中遵循的规则更加严格，并且会更加的客观，也就是说学术的客观性会比较多，在翻译中自己主观理解的加入可能较少。这里有一个例子。原文：“那晚上的天色不大好，…”刘先生译文：“The weather was not very good that night but，….”张先生译文：“It looked like rain that evening，….”而“天色不太好”并不代表就是要下雨了。虽然对于这一句的翻译并不影响整体，但是作为一个翻译来说，并不应该加入过多的理所当然的想象。 除了翻译家的职业经历会影响他们的翻译表现和态度，他们对于翻译的基本看法，也会影响翻译工作。刘先生认为翻译的最高境界是对原文韵味的再现。译作“韵味”就是原作的艺术内涵通过译文准确而富有文采的语言表达时所蕴含的艺术感染力，这能引起读者的美感共鸣。除此之外，刘先生还强调要把译文作为独立的文本来看待。具体说来就是对译文美感和韵味的展现虽是从词句入手的，同时还应注意译文包括内容和风格的整体效果。当两者有了矛盾，要变通前者来适应后者。从以上不难看出，刘先生在翻译时非常注重两方面，一是对原文韵味的保留，也就是原文内容中所蕴含的艺术吸引力的保留，另一方面就是对翻译整体效果的追求要超越对词语句子的效果的追求。 译文对比评析从哪些方面 匆匆英译文对比赏析 （1）匆匆 译文1：Transient Days 译文2：Rush （2）燕子去了，有再来的时候；杨柳枯了，有再青的时候；桃花谢了，有再开的时候。 译文1：If swallows go away, they will come back again. If willows wither, they will turn green again. If peach blossoms fade, they will flower again. 译文2：Swallows may have gone, but there is a time to return; willow trees may have died back, but there is a time of regreening; peach blossoms may have fallen, but they will bloom again. （3）但是，聪明的，你告诉我，我们的日子为什么一去不复返呢?——是有人偷了他们罢：那是谁?又藏在何处呢?是他们自己逃走了罢：现在又到了哪里呢? 译文1：But, tell me, you the wise, why should our days go by never to return? Perhaps they have been stolen by someone. But who could it be and where could he hide them? Perhaps they have just run away by themselves. But where could they be at the present moment? 译文2：Now, you the wise, tell me, why should our days leave us, never to return? ----If they had been stolen by someone, who could it be? Where could he hide them? If they had made the escape themselves, then where could they stay at the moment? （4）我不知道他们给了我多少日子；但我的手确乎是渐渐空虚了。译文1：I don’t know how many days I am entitled to 对《匆匆》的三个译本对比分析 翻译大家常说好的翻译作品要忠于原文，完全了解原文的写作背景，作者的写作风格；并把握理解作者表达的思想感情，挖掘作者原作中的内涵，捕捉原作里的神韵之美，以及丰富的感情。为了对比散文《匆匆》的三个不同译本我他特意查了一下《匆匆》的写作背景，然后读了几遍原文，深深地体味了一下朱自清的写作情感。顿时，感受到了朱自清对时间流逝的无奈焦急和惋惜之情。再读《匆匆》，似乎在听一首好听灵动的音乐，那朴素平淡的抒情气氛，优美的了律动美，委婉流畅悠远的的音律美油然而生。然后再欣赏三个不同的译本，我或多或少感受到了三个译本各自不同的特点。 乍眼一看，我特别喜欢张培基先生的译本，我觉得他译的真好。因为我看他的译文时，感受到了原作丰富的思想感情，浓厚的感染力，以及原文的美和意境。但是经过仔细品味，我发现其实三个译本各有千秋，不能单单说谁译的好与不好，只能说谁翻译的比较准确，更能符合原文之美，原文之意，原文之味。 以下是我对三个译本的对比分析欣赏：（朱纯深，张培基，张梦井先生的译本） 对比一：题目。 朱纯深译文：Rush 张培基译文：Transient Days 张梦井译文：Days Gone By Rush的意思是快速移动；急促；有仓促之意。我认为时间虽然短暂，但也不至于仓促过往，rush 没有美感，也不符合原文所表达匆匆之意，不太妥帖。 Transien t的意思是转瞬即逝的,短暂的(continuing for only a short time )把时间匆匆流逝，时间短暂之意表达的淋漓尽致，似乎让人觉得时间犹如过往云烟，捉不住，摸不着，与原文匆匆之意相符合，且具有美感。 Go b y的意思是流逝,过去（to past）多指时间的推移，不能表达匆匆之意。 综上所述，Transient Days更为贴切，而其他两个就欠妥帖。 对比二： 朱纯深：Swallows may have gone, but there is a time to return; willow trees may have died back, but there is a time of re-greening; peach blossoms may have fallen, but they will bloom again. 张培基：If swallows go away, they will come back again. If willows wither, they will turn green again. If peach blossoms fade, they will flower again. 对比功能分析与翻译 摘要：本文首先简要讨论了对比功能分析与翻译之间的密切联系。然后以一个实例指出，实际翻译过程是一个决策过程，对比功能分析可以为分析在翻译中可供选择的各种语言选项及其形式、语义特征和篇章、语用使用条件提供一个总体描述框架，因而对指导具体翻译和分析译文所作选择的得失具有很大的实用价值。 1. 引言 对比语言学的核心问题是研究方法问题。虽然从事语言对比的研究者很多，但是思考和提出系统的、专门用于语言对比的理论模式和框架的研究者却屈指可数。波兰格但斯克大学的Krzeszowski教授是其中的一位，而芬兰赫尔辛基大学的Chesterman教授是另一位这样的学者。如果说前者(1979, 1980)提出的“对比生成语法”模式(简介见许余龙1992: 180 -181; 2002: 163 -164)主要适用于句法结构对比的话，那么后者(1998)提出的“对比功能分析”模式的适用范围则要广泛得多，不仅可以用于词汇、形态、句法、语义、语用对比，也可以用于话语分析、文体、修辞和社会语言学对比(关于其研究方法以及在英汉对比中的应用，见许余龙2005)。而且，由于Chesterman是一位多语交际和翻译研究专家，出版过多部语言学和翻译研究方面的专著(如Chesterman, 1997; Chesterman & Wagner, 2002; Williams & Chesterman, 2002/2004)，因此他更加强调对比研究与翻译之间的密切联系。本文将从翻译和对比功能分析的性质和特点出发，讨论两者之间的密切联系，并以一个实例来说明对比功能分析对翻译的指导意义。 2. 对比功能分析与翻译之间的密切联系 2. 1翻译的性质和特点 著名捷克学者Levy (1967/1989: 38)指出，“从目的论的角度来说，翻译是一个交际过程：翻译的目的是将原文表达的知识信息传递给外国读者。而从译者在进行实际翻译工作时的任何一刻来看，翻译是一个决策过程：一系列的情景需要译者在一组(通常可明确定义的)选项中作出某种选择”。 翻译的这两个基本性质和特点在以后的一些研究中，特别是功能主义取向的研究中(如见Hatim, 1997/2001; Hatim &Mason, 1997; Nord, 1997/2001)，得到了进一步深入的探讨。也正是翻译的这两个基本性质和特点，使得语言文化之间的对比分析和翻译成为两种具有密切联系的研究。正如Hatim (1997/ 2001: 1)所指出，“要知道对比语言学是如何运作的，一种有效的方法是通过翻译；而要了解翻译的过程，一种有趣的方法是看译者在处理篇章时作出什么样的选择。” 2. 2对比功能分析的性质和特点 所谓对比功能分析，广义而言是从功能主义的角度来进行对比分析的一种研究方法。这一对比分析模式的特点是从察觉到的两种(或多种)语言可表达的相似意义出发，致力于确定这种相似的意义在不同的语言中是如何表达的，不同表达方式的句法、语义、篇章、语用、语境的使用条件是什么，在什么条件下会优先选用哪种形式，等等。对比功能分析与翻译之间的密切联系表现为：两者都旨在确定哪种形式在什么样的使用条件下是表达某一意义的最佳形式。而两者之间的区别则在于：翻译的任务是要确定，在某一特定使用条件下的一种源语语言形式所表达的意义，在目标语中的最佳表达形式是什么；而对比功能分析则致力于探 阐释学视角下《茵梦湖》三个中译本对比研究分析 随着翻译研究的不断深入，学者们从不同的角度，用不同的理论方法来探究与文学翻译相关的现象和问题。伽达默尔阐释学的三大哲学原则，即“理解的历史性”“视域融合”“效果历史”对文学翻译产生了很大影响。研究的文本来自德国著名诗意现实主义小说家特奥尔多·施笃姆的重要代表作品《茵梦湖》的三个中译本。通过对比分析郭沫若、巴金和杨武能三个中文译本，证明伽达默尔阐释学的三大哲学原则对文化翻译的指导价值和借鉴意义。 标签：阐释学三大哲学原则茵梦湖对比分析 一、引言 阐释学理论要追溯到“语言一历史传统”（Hohn，1998：91）这一翻译理念中，因为它自古希腊罗马时期就已经形成并被记录在Storig（1963）的选集中。它强调翻译活动的理解特点，并在文学翻译领域中体现得特别明显。阐释学翻译理论的基本假设之一是认识到两个文本之间没有等价关系。这被认为是语言翻译理论的第一次进步；另一方面，翻译也被理解为是一个决策过程，并确定译者的责任。对整部作品、作家以及写作动机理解的越多，对单一文本的理解就会越好。本文试图通过分析郭沫若、巴金和杨武能对德国中篇小说《茵梦湖》的三个不同译本，介绍伽达默尔诠释学的三个哲学原则及其对文学翻译的重要意义。 二、中篇小说《茵梦湖》以及伽达默尔的三大哲学阐释原则 （一）《茵梦湖》文本 《茵梦湖》是德国作家汉斯·台奥多尔·沃尔特森·施托姆创作的中篇小说，1849年首次出版问世，此后共出版30余次。通过这部小说的成功问世，施托姆在文学评论家和读者群中也迅速收获了知名小说家的美誉。这部中篇小说在19世纪下半叶仍然是他最著名的作品。小说源于一次回忆：一位名叫莱因哈德的老人在一个黑暗的小房间里思考了他年轻时的爱情；孩提时代，青梅竹马的伊丽莎白和他一起度过。为了外出求学，莱因哈德被迫离开了伊丽莎白。莱因哈德回来时，发现伊丽莎白变了。莱因哈德的学校朋友埃里克似乎对伊丽莎白表现出兴趣。莱因哈德再次离开，后来他通过母亲的书信获悉，伊丽莎白與埃里克已经结婚了。几年后，莱因哈德受邀来到茵梦湖畔探望这对已经拥有独立庄园的埃里希夫妇，但再次的重逢变得无法忍受，他决定永远离开伊丽莎白。 故事的背景发生在德国南部。这部小说在国内外都非常受读者欢迎，所以有关这部作品的研究也有很多。国内对《茵梦湖》的研究主要集中在以下两点：用一些文学理论来分析这部作品；借用不同的翻译理论来探究这部作品的中文译本。《茵梦湖》在国内比较受欢迎的译本有三个，本文基于郭沫若、巴金和杨武能的三种译本进行研究。 山东外语教学 Shandong Foreign Language Teaching Journal 2006年第4期(总第113期) 对比功能分析与翻译 许余龙 (上海外国语大学语言文学研究所,上海 200083) 收稿时间:2006206210 作者简介:许余龙(1950-),博士,教授,博士生导师。研究方向:对比语言学,英汉对比,篇章回指。 摘要:本文首先简要讨论了对比功能分析与翻译之间的密切联系。然后以一个实例指出,实际翻译过程是一个决策过程,对比功能分析可以为分析在翻译中可供选择的各种语言选项及其形式、语义特征和篇章、语用使用条件提供一个总体描述框架,因而对指导具体翻译和分析译文所作选择的得失具有很大的实用价值。 关键词:功能主义;对比分析;对比功能分析;翻译 中图分类号:H0 文献标识码:A 文章编号:100222643(2006)04200032061.0引言 对比语言学的核心问题是研究方法问题虽然从事语言对比的研究者很多,但是思考和提出系统的、专门用于语言对比的理论模式和框架的研究者却屈指可数。波兰格但斯克大学的K rzeszowski 教授是其中的一位,而芬兰赫尔辛基大学的Chesterman 教授是另一位这样的学者。如果说前者(1979,1980)提出的“对比生成语法”模式(简介见许余龙1992:180-181;2002:163-164)主要适用于句法结 构对比的话,那么后者(1998)提出的“对比功能分析”模式的适用范围则要广泛得多,不仅可以用于词汇、形态、句法、语义、语用对比,也可以用于话语分析、文体、修辞和社会语言学对比(关于其研究方法以及在英汉对比中的应用,见许余龙2005)。而且,由于Chesterman 是一位多语交际和翻译研究专家,出版过多部语言学和翻译研究方面的专著(如Ches 2terman ,1997;Chesterman &Wagner ,2002;Williams &Chesterman ,2002Π2004),因此他更加强调对比研究与 翻译之间的密切联系。本文将从翻译和对比功能分析的性质和特点出发,讨论两者之间的密切联系,并以一个实例来说明对比功能分析对翻译的指导意义。 2.0对比功能分析与翻译之间的密切联系2.1翻译的性质和特点 著名捷克学者Levy (1967Π1989:38)指出,“从目的论的角度来说,翻译是一个交际过程:翻译的目的是将原文表达的知识信息传递给外国读者。而从译者在进行实际翻译工作时的任何一刻来看,翻译是一个决策过程:一系列的情景需要译者在一组(通常可明确定义的)选项中作出某种选择”。 翻译的这两个基本性质和特点在以后的一些研究中,特别是功能主义取向的研究中(如见Hatim ,19972001;Hatim &Mas on ,1997;Nord ,1997Π2001),得 到了进一步深入的探讨。也正是翻译的这两个基本性质和特点,使得语言文化之间的对比分析和翻译成为两种具有密切联系的研究。正如Hatim (1997Π2001:1)所指出,“要知道对比语言学是如何运作的, 一种有效的方法是通过翻译;而要了解翻译的过程,一种有趣的方法是看译者在处理篇章时作出什么样的选择。” 2.2对比功能分析的性质和特点 所谓对比功能分析,广义而言是从功能主义的角度来进行对比分析的一种研究方法。这一对比分析模式的特点是从察觉到的两种(或多种)语言可表达的相似意义出发,致力于确定这种相似的意义在不同的语言中是如何表达的,不同表达方式的句法、语义、篇章、语用、语境的使用条件是什么,在什么条件下会优先选用哪种形式,等等。对比功能分析与 3 译文对比评析从哪些方面 第一作品主题。我们首先要明白这部作品或者文章的主题是什么。这就需要我们认真读下序言和简介，这是理解作品关键的地方。无论作品有多长，总有一个中心围绕着，我们要找出来。有时候，一部作品的主题非要看完才会明白。 第二作品人物。一部文学作品，一般会安排好几个主要人物。假如这部作品非常长，有好几部，那么人物就会很多，这与作者的设计有一定的关系。我们知道托马斯·哈代一生共写过十四部长篇小说，但是他的作品人物安排有一个特点就是：两位男主人公和一位女主人公，或者是两位女主人公和一位男主人公，这或与与他自己的体验有一定的关系。因此，我们了解作品的主要人物还是很重要的。 第三作品结构。一部好的文学作品，其中的结构安排会是比较细心的。很多鸿篇巨著，作者一般在结构上不会疏忽。这方面可以以巴尔扎克、雨果、司汤达为代表。说到结构我们会想到《一千零一夜》那种比较独特的结构，就是故事套故事的结构，这种结构后来还被很多文学作者所采用了。我们了解和熟悉作品结构，对于欣赏文学作品还是有着一定的意义。 第四作品风格。一部文学作品，一定有着比较独特的艺术风格，如浪漫主义、现实主义，古典主义，还是意识流。每部文学作品都有着不一样的艺术风格，每个时代的作者，一定会书写具有那个时代特征的作品。如果我们阅读十八世纪 的作品，或许会想到与启蒙运动相关的内容。此外，我们对作品的艺术风格有着一定的了解和领会，对欣赏文艺作品有一定的作用。 第五作品思想。当我们阅读一部文学作品的时候，总会被作者独特的思维和见解所吸引住，也许这就是其中的魅力所在。一部优秀的作品，其中的思想性一定与艺术性相匹配的。作者的人生态度积极或者消极都与那个时代的环境有着密切的关系。我们了解作品思想，实际上能够比较深入的理解那个时代的历史背景，这是分析一部文学作品不可缺少的。 《我为什么而活着》三种中译本的对比赏析 一．写作背景和原文内容介绍 《我为什么而活着》选自《贝特兰·罗素自传》，是这本自传的前言部分。这本书创作于作者晚年时期，它既是作者心灵的抒发，也是生命体验的总结。前言部分很好地总结了作者一生的精神生活：“三种单纯却又极强的激情支配了我的一生：对爱情的渴望，对知识的追求，和对人类苦难感到无法忍受的怜悯之情。”这三种激情是罗素一生在爱情、理智和道德三方面生活的动力。在理智生活中，罗素一生追求确实可靠的知识，然而在现实生活中，他又是一个极富感情的人。正是由于这种对爱情的热烈渴望，他才有了婚姻的波折，几次结合与离异。你可以不赞成他的行为，但却不能指责他感情虚伪。这便是爱情生活中的罗素。罗素以深刻的感悟和敏锐的目光，分析了人生中的三种激情，即对爱的渴望，对知识的追求和对人类苦难的同情。对爱的渴望，使人欣喜若狂，既能解除孤独，又能发现美好的未来。对知识的追求，使人理解人心，了解宇宙，掌握科学。爱和知识把人引向天堂般的境界，而对人类的同情之心又使人回到苦难深重的人间。作者认为这就是人生，值得为此再活一次的人生。 本文作为思考人生意义与价值的经典之作，被众人熟知。经过认真阅读三种不同的译本，并和原文进行对比比较，我认为这三种译本各有千秋。原文是一篇关于人生这样一个严肃话题的议论文--散文的一种，为语域中的正式语体。文章命题清楚，说理透彻，逻辑性强，段落严谨，用词讲究，风格凝重而又不乏诙谐。充满激情和生动的文字使文章在极具说服力的同时又具有文采，从这点看，三种译本都符合要求。 二．三种译本的对比赏析 首先从题目来看，第一种译本《我为什么生活》来源于《世界文学随笔精品大展》，作者泰云；第二种译本《三种激情》来源于《英国散文小书屋》，作者陈炼佳；第三种译本《我生活的目的》来自《英汉文体翻译教程》，作者陈新。纵观以上三种译本，三种激情四个字直接点明文章的中心思想，属意译，即是根据上下文译出来的。其他两种译文则属直译，直接根据题目译出来的，不管是我为什么生活还是我生活的目地，表述都不是那么得体，相较于这两种译文，我更倾向于三种激情。第一句话Three passions, simple but overwhelmingly strong, have governed my life，三种译本采取了不同的处理方法，泰译本采取了“合”的思想，直接把三个小短句合为一句话，译成三种单纯而极其强烈的激情支配着我的一生；陈炼佳译本中直接采取保留原有格式，用三个小短句来翻译，译成三种激情虽然简单，却异常强烈，它们统治着我的生命；而陈新译本中也采取了“合”的思想，把第一句翻译成在我的生活中起支配作用的有三种简单却又极为强烈的情感。第一句的处理上第一种比较好，既简明扼要又准确，第二句把governed译为统治，统治这个词带有感情色彩，而支配则比较中性，因此，支配要比统治恰当。在对第一段最后一句话The three passions, like great winds, have blown me hither and thither, in a wayward course, over a deep ocean of anguish, reaching to the very verge of despair的翻译中，陈新的版本更加表达了作者无奈和绝望的心情，这些情感像大风一样吹来吹去，方向不定，越过深沉痛苦的海洋，直达绝望的边缘。尽管原文里有have blown me 这样的字眼，陈新并没有像其他两位作者一样，《葛底斯堡演讲》三个中文译本的对比分析
译文对比分析
《傲慢与偏见》译文对比分析
《乡愁》两个英译本的对比分析
落花生两英译本的对比分析
译文对比评析从哪些方面
对《匆匆》的三个译本对比分析
对比功能分析与翻译
阐释学视角下《茵梦湖》三个中译本对比研究分析
对比功能分析与翻译
译文对比评析从哪些方面
《我为什么而活着》三种译本分析比较