JP2007187939A - English conversation learning apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an apparatus enabling a learner to efficiently learn English conversation. <P>SOLUTION: An individual-oriented English conversation learning apparatus successively displays an ordinary Japanese sentence, a Japanese sentence in the order of English words, image of the tense, and an English sentence; then makes along with the display, an English sentence by a native speaker generated, and allows a learner to learn the English conversation through imitation of the speech. In the apparatus, the content of speech by the learner is voice-recognized and graded and, according to the grading, the level of sentence to be displayed next is changed automatically. This learning course is conducted automatically, until the learning material comprising a plurality of sentences has been 100% assimilated. The level is set by controlling the time, during which the ordinary Japanese sentence, the Japanese sentence in English word order, and the tense image are displayed successively, and the time, until which the speech by a native speaker along with the display, is generated. Several tens to hundreds of sentences are prepared for a single learning material, and each sentence is repeated at least five times with the level being raised sequentially. The sentence is selected at random and displayed. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an apparatus for learning English conversation by a person whose native language is other than English, and relates to an English conversation learning apparatus that is automatically level-controlled by a computer program.

  Conventionally, as a means for learning English conversation, an English conversation school, an English teaching material provided on a CD, or an English correspondence education is generally used. However, the success or failure of acquiring English depends on the individual's sense of purpose and time / money, and there is a need for a means to learn English conversation more efficiently.

In response to such a request, for example, Patent Literature 1 discloses a language learning device using speech recognition.
According to the present invention, the pronunciation of the role selected by the learner is compared with the reference data and scored according to the degree of coincidence, and the appropriate next screen is automatically displayed according to the score, thereby improving learning efficiency. It becomes possible to make it.

  However, for example, the apparatus described in Patent Document 1 has a drawback that the learner needs to operate the apparatus one by one during the learning process, which makes it troublesome in the middle and reduces the willingness to learn.

  In addition, for Japanese whose native language is very different from English, such as Japanese, the difference in grammar is considered to be a barrier in learning English conversation. There was no English conversation learning device that provided a means of overcoming.

Also, unlike English, English often requires a tense, but there has been no conventional display of this tense in teaching materials.
In addition, the clearing rate (score) of the English conversation teaching materials is a display for enhancing the learner's willingness to learn, but there has been no conventional method for effectively displaying the learned clear rate in the learning process of the English conversation teaching materials. In other words, the existing teaching materials have a drawback that it is difficult to explicitly grasp both the mass and how much the learner has learned the teaching materials, and it is inevitably a self-judgment.
JP 2003-228279 A JP 7-168520 A Japanese Patent Laid-Open No. 2004-21102 Japanese Patent No. 2854652 Japanese Examined Patent Publication No. 7-11810

Therefore, in view of such a problem, an object of the present invention is to provide an individual correspondence learning apparatus that can learn English conversation more efficiently.
For this reason, the present invention recognizes the learner's pronunciation content by voice recognition and scores it, and automatically changes the level when the same sentence is displayed next time by the scoring, and the digestibility of the teaching material composed of a plurality of sentences is reduced. It is an object of the present invention to provide an English conversation learning tool in which the learning process automatically proceeds until reaching 100%.

  In order to achieve the above object, the English conversation learning apparatus of the present invention is controlled by a computer, and normally displays Japanese sentences, English-ordered Japanese sentences, tense images, and English sentences in sequence, and then generates native pronunciation of English sentences together with the display. In a personal English conversation learning apparatus that learns by imitating this, the learner's pronunciation content is recognized by voice recognition, and the level is automatically changed the next time the same sentence is displayed by the scoring. In addition, the learning process automatically proceeds until the digestibility of the teaching material composed of a plurality of sentences reaches 100%.

  Furthermore, the level is performed by controlling the time during which normal Japanese sentences, English-ordered Japanese sentences, tense images are sequentially displayed, and the time until the native pronunciation of English sentences is generated along with the display. To do.

  Furthermore, dozens to hundreds of the sentences are prepared in one teaching material, and each sentence is raised in order and repeated at least five times, or the sentence is selected and displayed at random. Features.

  In addition, there is a result file for index recording corresponding to the number of repetitions for each prepared sentence, and when it is determined that the randomly selected sentence is cleared by scoring, the sentence corresponds to the sentence in the result file. The index record is deleted, and the achievement rate of the teaching material at that time is evaluated based on the number of remaining index records in the result file.

  In addition, the level selected when the same sentence is displayed next time is increased by one level when the previous learning of the sentence is cleared, and when the previous learning of the sentence is not cleared, The number of sentence learning is increased by one.

Furthermore, the sentence level selected and displayed next by the scoring is one of the following five levels of patterns.
(1) Level 1 (Easy)
Japanese sentence (between / maximum) → English order Japanese sentence (between / maximum) → English sentence (between / maximum) → English pronunciation (2) Level 2
Japanese sentence (between / maximum) → English order Japanese sentence (between / maximum) → English thin characters (between / maximum) → English pronunciation (3) Level 3
Japanese sentences (between and middle) → English order Japanese sentences (between and middle) → English thinnest characters (between and middle) → English pronunciation (4) Level 4
Japanese sentence (between / small) → English order Japanese sentence (between / small) → …… (between / small) → English pronunciation (5) Level 5 (hardest)
Japanese sentence (between / large) → English sentence Japanese sentence thin (between / medium) → …… (between / small) → ……
(Note that “...” means that the display is omitted.)

Furthermore, the Japanese sentence in English order is characterized by the following two rules.
(1) Display in phrase units, and each is an expression that can be understood as Japanese, but when it is expressed in English, the part that should come after the phrase is enclosed in parentheses ().
(2) By displaying “-” as a pair, two distant clauses are connected to display the exact meaning.

Furthermore, when the scoring result is NG, the sentence is divided into phrase units, the series of learning is performed on sentences generated in a simpler sentence, and the tense image is composed of the following 15 patterns. It is characterized by.
・ Invariant events (no tense)
・ Past Events ・ Current Events ・ Future Guessing, Will ・ Current Completion, Experience, Continuation ・ Cause ・ Intended Future ・ Past Events Progress ・ Current Events Progress ・ Future Schedules, Plans ・ Past Continuous situation, state / repetitive situation in the past, state / current continuous situation, state / current repetitive situation, completion of state / past point, experience, continuation / reliable future

  The English conversation learning apparatus according to the present invention is controlled by a computer, recognizes the pronunciation content of a learner by voice, and scores it, and automatically changes the level when the same sentence is displayed next time based on the scoring, and includes a plurality of sentences. The learning process automatically progresses until the digestibility of the teaching material reaches 100%, so that the learner can experience the sentence and its clearness one after another in real time, and the interest can be maintained without getting bored with English conversation learning. Therefore, there is a remarkable effect that the learning effect is improved. On the other hand, a conventional learning device that requires an operation for each sentence has a drawback that the user's concentration is interrupted because the operation is troublesome, and the user immediately becomes disliked.

  In addition, the English conversation learning device of the present invention is controlled by a computer and normally displays Japanese sentences, Japanese words in order of Japanese words, tense images, and English sentences in sequence, then generates native pronunciation of the English sentences along with the display, and imitates this It is an individualized English conversation learning device that learns by doing, and by displaying a tense image, there is an effect that it is possible to intuitively understand the reaction to the tense that is particularly difficult for Japanese people.

  Furthermore, by preparing several dozens to several hundreds of the sentences in one teaching material, each sentence is raised in order and repeated at least five times, or the sentences are selected and displayed at random. , There is a change for each sentence, and there is a fresh taste than the one that repeats the same sentence monotonously.

  In addition, the effect of the sentence appearing one after another at random and the learner's speaking is close to the actual conversation situation (the conversation does not stop, the conversation is unpredictable) and is very natural in the sense of memorizing words. Is optimal. Furthermore, the memory of the same sentence can be ensured by providing an interval.

  Furthermore, the same sentence progresses to the upper level each time it is cleared, and if it cannot be cleared, it shifts to an easier level. Therefore, since the problem according to the ability of the learner is sequentially presented, the learning effect is high. In addition, since a problem suitable for one's level is presented, it is not possible to make it extremely or cannot be made extremely, and the learner does not get bored, so the motivation for learning increases.

  Also, there are index recording result files corresponding to the number of repetitions for each prepared sentence. When it is determined that the randomly selected sentence has been cleared by scoring, the index corresponding to the sentence in the result file. Since the record is deleted and the achievement rate of the teaching material at that time is evaluated based on the number of index records corresponding to the sentence in the result file, the progress of the teaching material is immediately known, and the learner's motivation is promoted. . Also, displaying the achievement rate on the screen has the effect of increasing the learner's willingness to achieve.

Furthermore, the Japanese sentence in English order is convenient for intuitively changing Japanese into English by providing the following two rules, and is associated with English sentences compared to the traditional Japanese sentence display. The effect is fast.
(1) Display in phrase units, and each is an expression that can be understood as Japanese, but when it is expressed in English, the part that should come after the phrase is enclosed in parentheses ().
(2) By displaying “-” as a pair, two distant clauses are connected to display the exact meaning.

  Further, when the scoring result is NG, the sentence is divided into phrase units, and the above-described series of learning is performed on sentences generated in a simpler sentence, so that a provision for a beginner is made. In this way, in addition to the normal level control, a simpler level learning process is prepared as a special case, so learning with simpler sentences is possible than with difficult sentences, and difficulty to clear Has the effect of mitigating.

  Hereinafter, the present invention will be described in detail based on examples with reference to the drawings.

  FIG. 1 shows an example of an English conversation learning apparatus according to this embodiment. This English conversation learning apparatus is realized by executing an English conversation learning program on a general personal computer 101.

  The monitor 102 displays an English conversation learning program screen based on the processing result of the personal computer 101. The keyboard 103 is used when inputting a user ID or the like. The mouse 104 is used when clicking various buttons on the screen. The microphone 105 is used when inputting a learner's pronunciation. The speaker 106 is used when outputting native English pronunciation and various notification sounds.

  FIG. 2 is a diagram illustrating a hardware configuration of the personal computer 101. A CPU 201 is a central processing unit and controls the entire personal computer 101. An HDD 202 is a hard disk drive, and stores an OS and an English conversation learning program of the present invention. A CD-ROM 203 is a CD-ROM drive and is used when an English conversation learning program provided by a CD-ROM medium is installed in the HDD 202. However, the present invention is not limited to this, and the English conversation learning program may be downloaded from the Internet and installed through a network interface (not shown). Alternatively, it may be provided as an ASP type service that does not require installation and provides functions through the Internet. Furthermore, it is preferable to be able to operate the English conversation learning device of the present invention on a portable terminal because English conversation can be learned in a game-like manner without choosing time and place.

  The memory 204 is a temporary storage device and is used for temporarily storing data processed by the CPU 201. An Input I / F 205 is an interface for an input device and is used to connect a keyboard 103, a mouse 104, and a microphone 105. An Output I / F 206 is an interface for an output device, and is used to connect the display 102 and the speaker 106. These hardwares are connected to each other by a bus 207.

Next, the operation of the English conversation learning apparatus of the present invention will be described using a screen diagram and a flow diagram. First, the login process will be described.
FIG. 3 is a diagram showing a login screen. When the English conversation learning apparatus of the present invention is activated, this screen is first displayed. The learner inputs a user ID in the user ID input field 301 and any speed level from 1 to 5 in the speed level input field 302. Incidentally, when the help button 303 is clicked, help is displayed, and when the end button 304 is clicked, the processing of the English conversation learning apparatus is ended. When the GO button 305 is pressed, login is performed based on the input user ID and speed level. Details of the login process will be described with reference to FIG.

  FIG. 4 is a flowchart showing details of the login process that is performed when the English conversation learning apparatus of the present invention is activated. As shown in FIG. 3, the English conversation learning apparatus first displays a login screen (S401). On this screen, the learner inputs the user ID and the speed level (S402). If the learner clicks the help button (YES in S403), the English conversation learning apparatus displays help (S404). If the learner clicks the GO button (YES in S405), the user ID and speed level are displayed. The following processing is performed.

  First, the English conversation learning apparatus holds a user ID in a variable UserName (S406) and sets a variable SpeedLevel from a speed level (S407). The Speed Level is 1.5 when the input speed level is “1”, 1.25 when “2”, 1.0 when “3”, 0.75 when “4”, “5”. In this case, 0.5 is set.

  Next, “UserName”. If the result, that is, UserName is “Hidemi”, it is determined whether there is a file “Hidemi.result” (S408). If this file does not exist (NO in S408), it means that the learner uses the English conversation learning apparatus for the first time, so Default. Copy the contents from the result file to “UserName”. A result file is created (S409). The “UserName”. result file, and Default. Details of the result file will be described later.

Next, the English conversation learning device initializes the speech recognition engine (S410). The speech recognition engine is a known program that can recognize a learner's pronunciation and evaluate it by a score or the like, and is incorporated in an English conversation learning apparatus. For example, the SRI International engine described in the following USP (US Patent Specification) is recommended as the speech recognition engine.
USP 5268990, USP 5816655, USP 5825978, USP 5864810, USP 6655498, USP 6266611.

Finally, the English conversation learning apparatus sets the phrase unit display flag to “Off” (S411). Details of the phrase unit display flag will be described later.
With the above processing, the English conversation learning device completes the login processing, and then shifts to the main processing. The main process will be described first from the screen transition, and the details of the process will be described later.

FIG. 5 shows a main screen displayed after the login process. First, the screen configuration of the English conversation learning apparatus will be described with reference to FIG.
The image display field 501 is a field for displaying an image related to English. The idiom display column 502 is a column for displaying idioms that are the theme of the English sentence. There are three guide lamps 503, a Ready lamp 503a, a Peak lamp 503b, and a Judge lamp 503c. The guide lamp 503 is used to indicate to the learner which stage is Ready, Speek, or Judge.

  The achievement rate display column 504 is a column for displaying the degree of clearing with respect to the total number of prepared English sentences. By displaying the achievement rate on the screen, the learner's motivation to achieve can be increased. In addition, when the achievement rate reaches 100%, a unique character string is displayed on the screen based on the logged-in character string, so the teacher determines that learning of the teaching material has been completed by reporting the character string It can be used in the same way, and it can be expected to be used in the field of education by using the effect of drastically reducing the teaching side's effort to grasp the progress of many students.

  The scoring result lamp 505 is used to indicate to the learner the results of the speech recognition engine recognizing and evaluating the pronunciation of the learner in five stages: Excellent, VeryGood, Good, NotGood, and Bad. By the way, if you become one of Excellent, VeryGood, or Good, you have cleared the English sentence that is the subject.

The English display column 506 is a column for displaying the English text that is the subject. The English order Japanese sentence display column 507 is a column for displaying a Japanese sentence in which the subject English sentence is translated as much as possible in order of the English sentence itself so that it can be fully understood in Japanese. For example, the English sentence is “I'd like to carry on this baggage with me.”
, Japanese sentences in English order,
“I want to bring it in with me.”
In English, the word that should be brought to the end of the clause in English is indicated by (), or the English text is “Where are you calling from?”
, Japanese sentences in English order,
"Where-(where are you calling)?-From"
In this way, the paired clauses are connected with-. The Japanese sentence displayed in the English order Japanese sentence display field 507 is prepared as content created in advance in this embodiment, but may be automatically generated from the English sentence that is the subject.

  In addition, Japanese sentences in English order are displayed without being bound by the meaning of idioms. This is because natives are not aware of idioms. For example, taking a preposition as an example, it is not remembered as a part of the idiom, but is used according to the image of the preposition.

  The tense image display field 508 matches the English text of the subject in the tense image diagram prepared in advance so that the viewpoint and background of the speaker that cannot be grasped from Japanese sentences alone can be understood. It is a column for displaying things.

  The effect of this tense image display is to convey to the learner the viewpoint, background, and nuance of the speaker that cannot be grasped in Japanese. This makes it possible to speak without memorizing English sentences.

  In this embodiment, the tense is specifically “invariant event (no tense)”, “past event”, “current event”, “future guess, will”, “current completion, experience, continuation”. , “Cause, intent with the future”, “progress of past events”, “progress of current events”, “future plans / plans”, “continuous past situations, states”, “repetitive past situations, states” ”,“ Current continuous status and status ”,“ current repetitive status and status ”,“ completion of past time points, experience and continuity ”, and“ certain future ”.

  The normal Japanese sentence display field 509 is a field for displaying the subject English sentence translated into correct Japanese. The sentence title menu 510 is used to select an English theme as a subject. The input volume 511 is for adjusting the volume of the microphone 105. The output volume 512 is for adjusting the volume of the speaker 106. The volume meter 513 is for displaying the volume input from the microphone 105.

  The start button 514 is for starting the main process, and the stop button 515 is for stopping the main process. Note that once the start button 514 is clicked, the main process is repeatedly executed without stopping until the stop button 515 is clicked or until all contents are completed.

As described with reference to FIG. 3, when the help button 303 is clicked, help is displayed, and when the end button 304 is clicked, the processing of the English conversation learning apparatus is ended.
The above is the screen configuration of the English conversation learning apparatus of the present embodiment. Next, basic screen transitions in the main process will be described with screen diagrams at each stage.

  FIG. 6 is a diagram showing a Ready stage screen. When the start button 514 is clicked on the screen of FIG. 5, the screen transitions to this screen. The Ready stage is a stage where the learner prepares before actually making a pronunciation, and the display is further divided into several stages. The details are as follows.

  First, a notification sound (chime) sounds and the Ready lamp 503a is turned on. Next, the contents are simultaneously displayed in the image display field 501, the idiom display field 502, the tense image display field 508, and the normal Japanese sentence display field 509. Next, the content is displayed in the English order Japanese sentence display field 507 after a short time, and in the English sentence display field 506 after a short time. Finally, after a short while, the native English pronunciation is output, and the Ready stage ends.

  FIG. 7 is a diagram showing a screen in the Peak stage. When the Ready stage described with reference to FIG. 6 ends, the screen transitions to this screen. The “Speak” stage is a stage where the learner actually pronounces the sound toward the microphone 105. First, a notification sound is emitted and the Peak lamp 503b is turned on, and “-Please Peak-” is displayed in the guide message column 701. Next, the learner pronounces the English text displayed in the English text display field 506 during a predetermined time limit. When the time limit has elapsed, the Peak stage ends.

  FIG. 8 is a diagram showing a screen waiting for the result of the Judge stage. When the Peak stage described with reference to FIG. 7 ends, the screen transitions to this screen. The Judge stage is a stage in which the pronunciation of the learner input in the Peak stage is evaluated and displayed using a speech recognition engine. The result waiting state shown in FIG. 8 is a state until the evaluation by the speech recognition engine is completed, the Judge lamp 503c is turned on, and “—>” indicating that the evaluation is in progress in the guide message column 701. A message is displayed.

  FIG. 9 is a diagram showing a screen of the result display state in the Judge stage. When the evaluation by the speech recognition engine is completed after the result waiting state shown in FIG. 8, the screen transitions to this screen. Here, based on the evaluation made by the speech recognition engine, the scoring result lamp 505 is turned on together with the notification sound (chime or buzzer), and if necessary, the necessary display is displayed in the English display column 506 and the guide message column 701. Do. For example, in the case shown in FIG. 9, the result is “Bad”, and the English display column 506 is highlighted such as changing the color of the English text that was evaluated poorly to red, and “* Advice such as “Clear word !!” is displayed. Then, this screen is displayed for a predetermined time, and the Judge stage ends.

  As described above, the main processing is one set in three stages of Ready, Peak, and Judge for one subject. When one assignment is completed, the process automatically proceeds to the next assignment, and this process is repeated until the learner clicks the stop button 515 or until all the contents are completed.

The Ready stage described with reference to FIG. 6 will be described here because other variations exist depending on the degree of difficulty.
The English conversation learning apparatus of the present embodiment prepares a display method of five patterns from level 1 to level 5 in the Ready stage, and this pattern can be changed depending on the degree of difficulty. As will be described later, an English sentence that is a subject is not cleared unless it is cleared at least five times, but the difficulty level is set to gradually increase during the five times.

The pattern already explained, that is, “First, the content is displayed at the same time in the normal Japanese text display field 509 etc., then in the English order Japanese text display field 507 at a short time, and then in English at a short time later. Each of the contents is displayed in the display field 506, and the native pronunciation of the English sentence is output at the end and after a while. "Is the least difficult of the five patterns. Here, this pattern is called “level 1”,
"Japanese (between / maximum) English pronunciation Japanese (between / maximum) English (between / maximum) English pronunciation"
It shall be expressed as

On the other hand, level 2 to level 5 can be expressed as follows.
Level 2
→ Japanese sentence (ma / dai) English order Japanese sentence (ma / dai) English thin character (ma / dai) English pronunciation level 3
→ Japanese (between and middle) English order Japanese (between and middle) English thinnest (between and middle) English pronunciation level 4
→ Japanese sentence (ma / small) English order Japanese sentence (ma / small)-(ma / small) English pronunciation level 5
→ Japanese sentence (between, large) English sentence Japanese character thin (between, middle)-(between, small)-

  That is, in the level 2 to the level 4, the “interval” is gradually reduced, and the English text is gradually displayed lightly and finally is not displayed. When level 5 is reached, English text is no longer displayed and English pronunciation is lost, and the learner answers only with Japanese text.

  In this way, the different display for each of the five levels is a mechanism that makes the learner aware of “English order”. In addition, by such training, it is possible to finally develop the simultaneous generation ability of “English thinking and speaking”.

The above is a variation of the Ready stage.
The screen transition of the main process has been described above. Next, details of the main process will be described.

First, a file used in the main process, Default. result and “UserName”. The result will be described. FIG. 10 shows Default. This shows the contents of the result file. This file is a text file, and in the example shown in the figure, five numbers from 1 to 10 are written, separated by commas. On the other hand, FIG. 11 shows “UserName”. re
An example of a sult file is shown. This file is also the Default. A text file similar to the result file, but with a Default. Compared to the result file, you can see that some numbers are reduced.

  The English conversation learning apparatus according to the present embodiment has “UserName”. The learning status of the learner is managed based on the result file. That is, the numbers shown in this file indicate the sentence numbers that need to be learned, and every time the assignment of the sentence number is cleared, the English conversation learning device deletes the corresponding numbers one by one. Specifically, in the example shown in FIG. 11, for example, three “1” numbers are deleted and there are only two, so the first sentence is cleared three times.

  On the other hand, if clearing of a subject of a certain sentence number fails, one number corresponding to the sentence number is added with the number of numbers being five as a limit. Specifically, in the example shown in FIG. 11, for example, if clearing fails in the first sentence, the number “1” increases from two to three.

  As described in the login process, “UserName”. When the learner uses the English conversation learning apparatus for the first time, the result. The result file is created by copying (S409 in FIG. 4). Default. Since each number is described in the result file by five, in order for the learner to completely clear the assignment of a certain sentence number, the sentence must be cleared at least five times.

  In addition, the achievement rate of the task is also “UserName”. Calculated using the result file. For example, in the example shown in FIG. 11, since the number originally 50 is reduced by 10 to 40, the achievement level is calculated as 20% from 10/50.

  Next, a sentence file in which sentences are recorded will be described. FIG. 12 shows the contents of the sentence file. In this sentence file, the sentence number 1201 is linked to the sentence title 1202, the idiom 1203, the English sentence 1204, the English sentence Japanese sentence 1205, the normal Japanese sentence 1206, the tense image No. 1207 are recorded.

  The sentence number 1201 is basically assigned every 10 as “10, 20, 30”, but there are some “21, 22, 23”. The sentence “21, 22, 23” is a sentence for learning by dividing the sentence of “20” for each phrase, and when the sentence of “20” cannot be cleared, the sentence is learned in more detail. It is prepared for.

  Next, the main process will be described in detail with reference to a flowchart. FIG. 13 is a flowchart showing the main process in detail. When the login process already described with reference to FIG. 4 is completed and the start button 514 is clicked on the displayed main screen (FIG. 5), this process is started.

  First, the English conversation learning device is “UserName”. One value is randomly obtained from the result array and stored as Gram_indx (S1301). Further, the number of Gram_indx is acquired and stored as Gram_cnt (S1302). Note that this Gram_cnt affects the difficulty level of the Ready stage of the “Ready, Peak, Judge process” later, specifically the weight process.

Next, the sentence number is set to a value obtained by multiplying (Gram_indx) by 10 and Gram_
It is stored as a number (S1303), the latest achievement rate is displayed (S1304), and "Ready, Peak, Judge processing" (S1305) is performed. This process will be described later with reference to FIG.

  Next, it is determined whether the phrase unit display flag is On or Off (S1306). If it is Off, the scoring result is determined (S1307). Here, if the scoring result is one of Excellent, VeryGood, Good, and it is determined to be OK (OK in S1307), “UserName”. One value of Gram_indx is deleted from the result array. If the achievement rate is not 100% (NO in S1309), the process is repeated again from the beginning (S1301).

  On the other hand, if the scoring result is either NotGood or Bad and it is determined as NG (NG in S1307), it is confirmed that Gram_cnt is less than 5, that is, there are not five or more Gram_indx. One value of Gram_indx is added to the result array (S1310).

  If there is a phrase unit sentence in the sentence whose score is NG (S1311), the phrase unit display flag is set to On (S1312), 1 is added to Gram_number (S1313), and Gram_cnt is set to 0. (S1314). That is, the phrase unit display flag is a flag for managing whether or not to display the phrase unit sentence. The reason why 1 is added to Gram_number is that, for example, the sentence unit sentence of the 20th sentence is recorded in the 21st, 22nd and 23rd sentences. The reason why Gram_cnt is set to 0 is to control “Ready, Speek, Judge processing” as described above.

  When the phrase unit display flag becomes On and the “ready, spike, and judge processing” is completed using the phrase unit sentence, it is next determined whether there is a sentence having a number obtained by adding 1 to Gram_number, that is, the next phrase unit sentence. (S1315). If this exists, 1 is further added to Gram_number (S1316), and the next sentence unit sentence is selected. On the other hand, if there is no following phrase unit sentence ("No" in S1315), it means that the phrase unit sentence has been learned, so the phrase unit display flag is set to Off (S1317), and the scoring result is OK. The same processing as in (1308, S1309) is performed.

  Next, the details of the “Ready, Peak, Judge process” will be described. FIG. 14 is a flowchart showing details of the “Ready, Peak, Judge process” shown in S1305 of FIG.

  First, the sentence information of Gram_number is acquired from the sentence file as shown in FIG. 12 (S1401). Next, the Ready lamp is turned on to output a notification sound (S1402), and images, idioms, tense images, and normal Japanese sentences are displayed. Note that the image is Gram_number. The tense image is a tense image No. Items corresponding to 1207 are displayed.

Next, after weight processing is performed (S1404), English sentences in Japanese are displayed (S1405). The wait time for this wait process is
Wait time = (2 seconds when Gram_cnt is 0, 1 second at 1 time, 2 seconds at 1 hour, 2 seconds at 3 hours, 3 seconds at 4 hours, 4 seconds at 5 times) × SpeedLevel
The value depends on the number of cleared sentences and the speed level entered on the login screen. In addition, Japanese sentences in English order are displayed in thin characters depending on the difficulty level.

Next, after performing the weight process again (S1406), an English sentence is displayed (S1407). The wait time for this wait process is
Wait time = (2 seconds when Gram_cnt is 0, 1 second at 1 time, 2 seconds at 2 time, 2 seconds at 3 hours, 3 seconds at 4 hours, 4 seconds at 5 times) × SpeedLevel
It is calculated by the formula. Also, English text may not be displayed depending on the difficulty level, or may be displayed in thin characters.

Next, after further weight processing (S1408), English native pronunciation is output (S1409). The wait time for this wait process is
Wait time = (4 seconds when Gram_cnt is 0, 1 second at 1 time, 1 second at 2 time, 2 seconds at 3 time, 3 seconds at 4 time, 4 seconds at 5 time) × SpeedLevel
It is calculated by the formula. English native pronunciation is Gram_number. Outputs a prepared voice file such as wav. However, depending on the level of difficulty, this English native pronunciation may not be output.

  When the processing in the Ready stage as described above is completed, next, the Peak lamp is turned on to output a notification sound (S1410), and the learner's pronunciation is accepted (S1411). The learner's pronunciation is passed to the recognition engine for evaluation. Then, the Judge lamp is turned on to output a notification sound (S1412), the result is obtained from the recognition engine, and the scoring result is displayed (S1413). Based on the OK / NG result obtained here, the processing described with reference to FIG. 13 is performed.

The Japanese sentence in English order has the following two rules.
(1) In English, the word or sentence existing at the end of the phrase is displayed in parentheses and displayed before the phrase.
(2) By displaying “-” as a pair, two distant clauses are connected to display the exact meaning.
The effect of Japanese text in parentheses is as follows.

When mapping English sentences with different grammar and Japanese sentences, it is generally accepted that it is necessary to first divide them into understandable phrase units.
However, there are two ways of expressing the phrase unit in Japanese: English word order or Japanese word order.
Simple example) This is a pen.
This is a pen (English word order)
This is a pen. (Japanese word order)

As can be seen from this example, it is difficult to understand the meaning in the English word order, and it is necessary to always be aware of the assembly in the English order in the Japanese word order. Therefore, as a result of studying the thought path when the native actually thinks and speaks at the same time, it was found that this example is as follows.
"Conscious of one pen, this is that one pen"

In other words, what comes behind in English word order already exists in consciousness, and in terms of understandable phrases, it is close to the Japanese word order. Therefore, by expressing this in Japanese as follows, it is possible to acquire a native thought utterance path. This has already been proved by experiments.
"This is (one pen)."

  According to the present invention, it is very efficient as a personal correspondence type English conversation learning apparatus, and it is long lasting. Therefore, the learning effect is remarkable and the industrial utility as a teaching tool is high.

It is the figure which showed an example of the English conversation learning apparatus in a present Example. 2 is a diagram illustrating a hardware configuration of a personal computer 101. FIG. It is the figure which showed the login screen. It is the flowchart which showed the detail of the login process implemented when starting the English conversation learning apparatus of this invention. It is the figure which showed the main screen displayed after a login process. It is the figure which showed the screen of a Ready stage. It is the figure which showed the screen of the Peak stage. It is the figure which showed the screen of the waiting state for the result of a Judge stage. It is the figure which showed the screen of the result display state of a Judge stage. Default. It is the figure which showed the contents of the result file. “UserName” for learners who have learned to some extent. It is the figure which showed an example of a result file. It is the figure which showed the content of the sentence file. It is a flowchart showing the main process in detail. FIG. 14 is a flowchart showing details of “Ready, Peak, Judge processing” shown in S1305 of FIG.

Explanation of symbols

101 PC
102 Monitor 103 Keyboard 104 Mouse 105 Microphone 106 Speaker 201 CPU
202 HDD
203 CD-ROM
204 Memory 205 Input I / F
206 Output I / F
207 Bus 303 Help button 304 End button 503 Guide lamp 505 Scoring result lamp 514 Start button 515 Stop button

Claims (10)

A personalized English conversation that is controlled by a computer, and displays normal Japanese sentences, English-ordered Japanese sentences, tense images, and English sentences in sequence, and then generates native pronunciation of the English sentences along with the display and imitates the pronunciation. In the learning device,
The learner's pronunciation is recognized and scored, and the level is automatically changed the next time the same sentence is displayed by the scoring, and the learning is continued until the digestibility of the teaching material consisting of a plurality of sentences reaches 100%. An English conversation learning device characterized in that the course proceeds automatically.   The level is performed by controlling the time during which normal Japanese text, English word order Japanese text, tense images are sequentially displayed and the time until the native pronunciation of the English text is generated with the display. The English conversation learning apparatus according to claim 1.   3. The English conversation learning apparatus according to claim 1, wherein several tens to several hundreds of the sentences are prepared for one teaching material, and each sentence is sequentially raised in level and repeated at least five times.   4. The English conversation learning apparatus according to claim 1, wherein the sentence is selected and displayed at random.   There are index record result files for the number of repetitions for each prepared sentence, and when it is determined that the randomly selected sentence is cleared by scoring, the index record corresponding to the sentence in the result file is recorded. 5. The English conversation learning apparatus according to claim 1, wherein the achievement rate of the teaching material at that time is evaluated based on the number of remaining index records in the result file.   When the same sentence is displayed next time, the level selected when the previous learning of the sentence is cleared is increased by one step, and when the previous learning of the sentence is not cleared, the level of the sentence is increased. 6. The English conversation learning apparatus according to claim 1, wherein the number of learning is increased by one. 7. The English conversation learning apparatus according to claim 1, wherein a sentence level selected and displayed next by the scoring is one of the following five levels of patterns.
(1) Level 1 (Easy)
Japanese sentence (between / maximum) → English order Japanese sentence (between / maximum) → English sentence (between / maximum) → English pronunciation (2) Level 2
Japanese sentence (between / maximum) → English order Japanese sentence (between / maximum) → English thin characters (between / maximum) → English pronunciation (3) Level 3
Japanese sentence (between and middle) → English order Japanese sentence (between and middle) → English thinnest character (between and middle) → English pronunciation (4) Level 4
Japanese sentence (between / small) → English order Japanese sentence (between / small) → …… (between / small) → English pronunciation (5) Level 5 (hardest)
Japanese sentence (between / large) → English sentence Japanese sentence thin (between / medium) → …… (between / small) → ……
(Note that “...” means that the display is omitted.) The English conversation learning apparatus according to any one of claims 1 to 7, wherein the Japanese sentence in the English order is provided with the following two rules.
(1) Display in phrase units, and each is an expression that can be understood as Japanese, but when it is expressed in English, the part that should come after the phrase is enclosed in parentheses (). (2) By displaying “-” as a pair, two distant clauses are connected to display the exact meaning.   9. The method according to claim 1, wherein when the scoring result is NG, the sentence is decomposed into phrase units, and the series of learning is performed on the sentence generated in a simpler sentence. The English conversation learning apparatus according to Item 1. The English conversation learning apparatus according to claim 1, wherein the tense image includes the following 15 patterns.
・ Invariant events (no tense)
・ Past event ・ Current event ・ Future speculation, will ・ Current completion, experience, continuation ・ Cause ・ Intended future ・ Past event progress ・ Current event progress ・ Future schedule, plan ・ Past Continuous situation, state / repetitive situation, state / current continuous situation, state / current repetitive situation, completion of state / past, experience, continuation / reliable future