gtg - Generative Tree Gadget

<HTML> <HEAD> <meta http-equiv="content-type" content="text-html; charset=utf-8"> <TITLE>gtg - Generative Tree Gadget</TITLE> <style> h2 {color:#A60000;} h3 {color:#C08700;} h4 {color:#C08700;} h5 {color:#C08700;} h6 {color:#C08700;} tt {color:#A60000; font-weight:bold; font-family:"Gentium";} </style> </HEAD> <BODY BGCOLOR="#FFFFFF" TEXT="#000000"> <form action="" name="theform"> <script language="javascript" type="text/javascript"> // Script (C) 2018 by Mark Rosenfelder. // Use Lexicon to build argument structure // Morphology should be done in some easy way // Build VP struture, use xforms to 'make it real' var theform; var lex; var psr; var showpsr = false; var fullform = false; var embedding = false; var inf = false; // Random int from 0 to n function randomInt(max) { return Math.floor(Math.random() * max) ; } // Return the word portion of a lexeme (before the :) function theword(s) { if (s == null || s == "") return ""; var i = s.indexOf(':'); if (i == -1) return s; return s.substr(0,i); } // Return the category of a lexeme (just after the :) function thehead(s) { if (s == null || s == "") return ""; var spp = s.split(":") if (spp.length < 2) return s; return spp[1]; } // Return the third parameter of a lexical item function thearg(s) { var spp = s.split(":") if (spp.length < 2) return ""; return spp[2]; } // Create a node in the tree function Node(h, left, mid, right) { var node = {head:"X", left:"", mid:"", right:""}; node.head = h; node.left = left; node.mid = mid; node.right = right; } // Is this a left bracket? function Bra(t) { return t.charAt(0) == '['; } // Is this a right bracket? function Ket(t) { return t == "]"; } // General function: replace the word in a w[] entry. // That is, replace just the part up to the first : function ReplaceWord(wd, tgt) { var colon = wd.indexOf(':'); return tgt + wd.substr(colon); } // Similar, but replaces parameter n // If n is 0, use ReplaceWord, it's faster function ReplaceParm(wd, n, tgt) { var partz = wd.split(":"); var s = ""; for (var i = 0; i < partz.length; i++) { if (s != "") s += ":"; if (i == n) s += tgt; else s += partz[i]; } return s; } // Given a verb and the desired form, change the verb to that form. // Example: // tgt = Prog // sit:V:PP:s sits Past sat Prog sitting Perf sat // change "sit" to "sitting" function Inflect(wd, tgt) { var vpp = wd.split(":"); if (vpp.length < 4) return wd; // STD this should probly be in xform lg instead // For N/Pron, insert number as well if (vpp[1] == "N" || vpp[1] == "Pron") tgt = theNumber(wd) + tgt; var forms = vpp[3].split(" "); for (var i = 0; i < forms.length; i += 2) { if (forms[i] == tgt) return ReplaceWord(wd, forms[i+1]); } return wd; } // Does this space-separated list contain element tgt or tgt2? function Includes(s, tgt, tgt2) { var spp = s.split(" "); for (var i = 0; i < spp.length; i++) { if (spp[i] == tgt || spp[i] == tgt2) return true; } return false; } var t; var o = null; var oo = null; var w = null; // Apply a rule // If rule doesn't apply, this will do nothing function Apply(r) { oo = null; var partz = r.split("="); if (partz.length < 2) return false; var doMe = new Array(o.size); var doSome = false; // Mark elements that are selected by this rule for (var i = 0; i < o.length; i++) { var word = o[i]; doMe[i] = theword(word) == partz[0]; if (doMe[i]) doSome = true; } if (!doSome) return false; // Some rules only apply about 1/3 the time if (partz.length > 2) if (partz[2] == "?") if (Math.random() < 0.667) return false; // Expand target(s) oo = new Array(); for (var i = 0; i < o.length;i++) { if (doMe[i]) { // This one matches the target; expand it oo.push("[" + partz[0]); var expans = partz[1]; var pipe = expans.indexOf('|'); if (pipe != -1) { // Multiple choice. Bias toward first one var altz = expans.split("|"); if (Math.random() < 0.33) expans = altz[0]; else { expans = altz[randomInt(altz.length)]; // Embed sentences only one level deep if (embedding && Includes(expans, "S", "Sinf")) expans = altz[0]; } } compz = expans.split(" "); for (var j = 0; j < compz.length; j++) { var tgt = compz[j]; if (tgt.charAt(0) == '?') { // ?X means sometimes insert X if (Math.random() < 0.33) oo.push(tgt.substr(1)); } else if (tgt.charAt(0) == '(') { // (X) means insert 0 to 2 copies of X tgt = tgt.substr(1, tgt.length - 2); var r3 = randomInt(3); for (var k = 0; k < r3; k++) oo.push(tgt); } else { oo.push(tgt); } } oo.push("]"); } else { oo.push(o[i]); } } return true; } // Does element o[i] or w[i] match str? // Elements syntax: X is a category; :X is a word; X:Y is a word X of category X // For an XP, write e.g. [NP function Match(i, str) { var cat = str; var word = ""; var colon = str.indexOf(':'); if (colon != -1) { cat = str.substr(0,colon); word = str.substr(colon+1); } var catmatch = cat == "" || theword(o[i]) == cat; var wordsame = word == "" || theword(w[i]) == word; return catmatch && wordsame; } var matches = null; // Find the sequence str within our sentence (o and w) // Str can be more than one element // Return the location within the o array, or -1 if not found function OFind(o, str) { var tgt = str.split(" "); matches = new Array(tgt.length + 1); for (var i = 0; i < o.length; i++) { var ok = false; var iOrig = i; for (var j = 0; j < tgt.length && i < o.length; j++) { ok = Match(i, tgt[j]); if (!ok) break; matches[j] = i; if (Bra(tgt[j])) { // This is an XP; skip to its end var level = 0; for (; i < o.length; i++) { if (Bra(o[i])) level++; else if (Ket(o[i])) level--; if (level == 0) break; } } i++; } matches[tgt.length] = i; i = iOrig; if (ok) return i; } return -1; } // Apply a transformation function Transform(ins, outs) { if (OFind(o, ins) == -1) { t += " - Target not found"; return false; } t += " - " + ins + " > " + outs; oo = new Array(); var ww = new Array(); var remloc = matches[matches.length - 1]; for (var i = 0; i < matches[0]; i++) { oo.push(o[i]); ww.push(w[i]); } var orig = ins.split(" "); var repz = outs.split(" "); for (var i = 0; i < repz.length; i++) { var el = repz[i]; var el0 = el.charAt(0); if (el0 == '»') { // Lexicon lookup var elz = el.substr(1).split(":"); var wd = Lookup(elz[0]); if (wd != "") { if (elz.length == 1) oo.push(thehead(wd)); else oo.push(elz[1]); ww.push(wd); } } else if ('0123456789'.indexOf(el0) == -1) { // Category oo.push(el); ww.push(""); } else { var r = parseInt(el); // Element from input string (by number) if (r >= 0 && r < orig.length) { for (var j = matches[r]; j < matches[r+1]; j++) { var wd = w[j]; oo.push(o[j]); ww.push(wd); } } } } for (var i = remloc; i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } o = oo; w = ww; return true; } // Return the clausemates of this V // STD: decide what direction to go - this could be a checkbox function VP(o, i) { var level = 0; var u = ""; for (var j = i + 1; j < o.length; j++) { var t = o[j]; if (Bra(t)) { if (level == 0) u += " " + t.substr(1); level++; } else if (Ket(t)) { level--; } else if (level == 0) u += " " + t; } if (u != "") u = u.substr(1); return u; } // Inflect word for plural function Pluralize(wd) { wd = Inflect(wd, "p"); // Add to arg so we know we changed this wd = ReplaceParm(wd, 2, thearg(wd) + " p"); return wd; } // Select lexical items function Lexicalize(o) { var w = new Array(o.length); // For each word in the sentence for (var i = 0; i < o.length; i++) { w[i] = ""; // Brackets don't get words if (Bra(o[i]) || Ket(o[i])) { continue; } // Go through lexicon var ll = new Array(); for (var j = 0; j < lex.length; j++) { var wd = lex[j]; if (o[i] == thehead(wd)) { if (o[i] == "Comp" && inf) { wd = "inf:Comp"; } if ((o[i] == "N" || o[i] == "Pron") && Math.random() < 0.33) { // Select plural for nouns, sometimes wd = Pluralize(wd); } if (o[i] == "V") { // Verbs must match # of arguments var vp = VP(o, i); if (vp != thearg(wd)) continue; } ll.push(wd); } } if (ll.length == 0) continue; // Select one randomly var m = randomInt(ll.length); w[i] = ll[m]; if (w[i] == null) t += "<br/>NULL"; } return w; } // Output tree as unreadable string // Set fullform to output entire word rather than 1st section function Stringize(o, w) { var s = ""; var level = 0; for (var i = 0; i < o.length; i++) { var t = o[i]; if (Bra(t)) { level++; var c = "Purple"; if (t == "[NP") c = "red"; else if (t == "[PP") c = "orange"; else if (t == "[S") c = "blue"; else if (t == "[TP") c = "green"; s += '<span style="color:' + c + '">'; } s += t + " "; if (Ket(t)) { level--; s += "</span>"; } if (w != null && w[i] != null) { s += '<span style="color:black;font-weight:bold;">'; if (fullform) s += w[i]; else s += theword(w[i]); s += '</span> '; } } return s; } // Output tree as even more unreadable string for phpSyntaxTree // Should be same as Stringize but: no colors; brackets round words; spaces > %20 function wordbrax(o, w) { var s = ""; for (var i = 0; i < o.length; i++) { if (Bra(o[i]) || Ket(o[i]) || w[i] == null) s += o[i]; else s += "[" + o[i] + "%20" + w[i] + "]"; } return s; } // Output pronounced words only function Pronounce(w) { var s = ""; var punct = "."; for (var i = 0; i < w.length; i++) { if (w[i] == null || w[i] == "") continue; var wd = theword(w[i]); if (wd == "Q") punct = "?"; else if (wd != "ø" && wd != "inf" && wd != "past" && wd != "pres") { if (s != "") s += " "; s += wd; } } s += punct; return s.italics(); } // Find the first element of category cat function FindFirst(cat) { var catz = cat.split("|"); for (var i = 0; i < o.length; i++) { for (var c = 0; c < catz.length; c++) if (theword(o[i]) == catz[c]) return i; } return -1; } // Find the first element of word w function FindFirstWord(tgt) { for (var i = 0; i < o.length; i++) { if (w[i] != null && theword(w[i]) == tgt) return i; } return -1; } // Insert a word just at a location in both o[] and w[] function Insert(tgt, wd, cat) { w.splice( tgt, 0, wd); o.splice( tgt, 0, cat); } // Find first item of class c from position i // c can list multiple classes separated by | // special class Aux is in third position function FindNext(c, i, dir) { if (i == -1) i = 0; var incr = 1; var iOrig = i; if (dir == "-") incr = -1; var cc = c.split("|"); // For each word in the indicated direction for (i += incr; i >= 0 && i < w.length; i += incr) { for (var j = 0; j < cc.length; j++) { if (o[i] == cc[j] || thehead(w[i]) == cc[j]) return i; } } // ±: if + direction failed, try - if (dir == "±") return FindNext(c, iOrig, "-"); return -1; } // Find the next word function NextWord(i) { for (i++; i < w.length; i++) { if (w[i] != null && w[i] != "" && theword(w[i]) != "ø") return i; } return -1; } // Find the last element of this constituent function FindEnd(tgt) { var level = 0; for (var i = tgt + 1; i < o.length; i++) { if (Bra(o[i])) level++; else if (Ket(o[i])) { level--; if (level < 0) return i + 1; } } return o.length; } // Flip two constituents function Flip(t1b, t2b) { if (t2b < t1b) { var temp = t2b; t2b = t1b; t1b = temp; } var t1e = FindEnd(t1b); var t2e = FindEnd(t2b); oo = new Array(); ww = new Array(); // Portion before t1 is copied as is var i = 0; for (; i < t1b && i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } // Copy t2 into t1's spot for (i = t2b; i < t2e && i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } // Portion in between is copied as is for (i = t1e; i < t2b && i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } // Copy t1 into t2's spot for (i = t1b; i < t1e && i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } // Portion after t2 is copied as is for (i = t2e; i < o.length; i++) { oo.push(o[i]); ww.push(w[i]); } o = oo; w = ww; } // Lexicon lookup function Lookup(item) { for (var l = 0; l < lex.length; l++) { if (theword(lex[l]) == item) return lex[l]; } return ""; } // Safe lookahead function Look(partz, i) { if (i + 1 < partz.length) return partz[i + 1]; else return ""; } var rcheck = false; // Apply a rule, e.g. // if Prog find V inflect ing function Xform(x) { var partz = x.split(" "); var tgt = -1; var clip = ""; var loop = -1; var nloop = 0; for (var i = 0; i < partz.length; i++) { var cmd = partz[i]; if (cmd == "sub") { if (!embedding) return false; } if (cmd == "main") { if (embedding) return false; } if (cmd == "maybe") { // Optional transformations happen at a fixed 1/3 rate if (Math.random() < 0.667) return false; } if (cmd == "checkif") { var ins = ""; for (i = i + 1; i < partz.length; i++) { if (ins != "") ins += " "; ins += partz[i]; } rcheck = OFind(o, ins) != -1; t += " check is " + rcheck; return; } if (cmd == "yes") { if (!rcheck) return; } if (cmd == "no") { if (rcheck) return; } if (cmd == "start") { // Remember where we are loop = i; } if (cmd == "loop") { // Loop back to remembered location nloop++; if (nloop < 10) i = loop; } if (cmd == "if") { // Existence check var want = true; if (Look(partz, i) == "no") { want = false; i++; } var item = Look(partz, i); i++; if (item == "word") { item = Look(partz,i); i++; tgt = FindFirstWord(item); } else tgt = FindFirst(item); t += " searching " + item + " " + tgt; // If that failed, exit if (want && tgt == -1) return; if (!want && tgt != -1) return; continue; } if (cmd == "find") { // Find an item; save it as the target // STD word var item = Look(partz, i); i++; var start = 0; var dir = "+"; if (item == "next") { start = tgt; item = Look(partz, i); i++; } else if (item == "previous") { start = tgt; item = Look(partz, i); i++; dir = "-"; } tgt = FindNext(item, start, dir); t += " finding " + tgt; if (tgt == -1) return; continue; } if (cmd == "object") { // Find the sentence's object (its [NP) tgt = FindNext("[NP", o.length, "-"); if (tgt < FindNext("[VP", o.length, "-")) tgt = -1; // we found the subject if (tgt == -1) return; } if (cmd == "subject") { // Find the sentence's subject (its [NP) tgt = FindNext("[NP", 0, "+"); if (tgt == -1) return; } if (cmd == "ifword" && tgt != -1) { // Is the next actual word equal to the one specified? while (tgt < o.length && theword(w[tgt]) == "") tgt++; var item = Look(partz, i); i++; t += " ifword " + item + " " + theword(w[tgt]); if (item != theword(w[tgt])) return; } if (cmd == "inflect" && tgt != -1) { // Inflect current word with the given inflection var afx = Look(partz, i); i++; if (afx == "clip") afx = clip; t += " inflecting " + theword(w[tgt]) + "+" + afx; w[tgt] = Inflect(w[tgt], afx); continue; } if (cmd == "transform") { // Transform one structure into another var ins = ""; for (i = i + 1; i < partz.length; i++) { if (partz[i] == "into") break; if (ins != "") ins += " "; ins += partz[i]; } var outs = ""; for (i = i + 1; i < partz.length; i++) { if (partz[i] == "end") break; if (outs != "") outs += " "; outs += partz[i]; } Transform(ins, outs); } if (cmd == "copy" && tgt != -1) { // Copy word information to clipboard if (Look(partz, i) == "num") { i++; clip = theNumber(w[tgt]); } else { clip = w[tgt]; if (clip == "") clip = "#" + tgt; } } if (cmd == "cut" && tgt != -1) { // Copy plus zero out old contents clip = w[tgt]; w[tgt] = "ø"; } if (cmd == "paste" && tgt != -1 && clip != "") { // Paste clipboard to target word t += " paste to " + theword(w[tgt]); if (w[tgt] == null || w[tgt] == "" || w[tgt] == "ø") w[tgt] = clip; else { if (clip == "s" || clip == "p") { w[tgt] = Pluralize(w[tgt]); } else if (o[tgt] == "T") { w[tgt] = theword(w[tgt]) + "+" + clip; } else w[tgt] = clip; } } if (cmd == "lex") { // Look up a lexical item; add it to the clipboard var item = Look(partz, i); i++; clip = Lookup(item); } if (cmd == "cuttense" && tgt != -1) { // Copy tense from this word to the clipboard var ew = theword(w[tgt]); var plus = ew.indexOf("+"); if (plus == -1) { clip = ew; w[tgt] = "ø"; } else { clip = ew.substr(0, plus); var rem = ew.substr(plus+1); w[tgt] = ReplaceWord(w[tgt], rem); } t += " copying tense " + clip; } if (cmd == "numclip") { // Add subject number to clipboard clip += FindNumber(); } if (cmd == "insert" && tgt != -1 && clip != "") { // Insert clipboard as new word after current location var cat = Look(partz, i); i++; if (cat == "") cat = thehead(clip); t += " inserting " + theword(clip); Insert(tgt + 1, clip, cat); } if (cmd == "append" && clip != "") { // Insert clipboard as new word at end t += " appending " + theword(clip); o.push(thehead(clip)); w.push(clip); } if (cmd == "prepend" && clip != "") { // Insert clipboard as new word at beginning, i.e., just after [S t += " prepending " + theword(clip); Insert(1, clip, thehead(clip)) } if (cmd == "flip" && clip != "" && tgt != "") { // Switch two constituents // STD word level not done if (clip.charAt(0) == '#') { var n = parseInt(clip.substr(1)); t += " flipping " + n + " and " + tgt; Flip(n, tgt); } } } } // Given a word, return its number s/p // If it's not defined, return "" function theNumber(wd) { var arg = thearg(wd); args = arg.split(" "); for (var i = 0; i < args.length; i++) { if (args[i] == "s" || args[i] == "p") return args[i]; } return ""; } // Find the inflection of the subject, which is stored in thearg of the word function FindNumber() { // Find the inflection of the subject, if any var n = FindFirst("N|Pron"); t += " findnumber " + n + " - " + thearg(w[n]); if (n != -1) return theNumber(w[n]); return ""; } // Used by DoDefine. Apply ins > outs for each sep-separated argument in s. function DoArgs(s, sep, ins, outs) { var args = s.split(sep); var s = ""; for (var j = 0; j < args.length; j++) { if (s != "") s += sep; if (args[j] == ins) s += outs; else { if (sep == " " && args[j].indexOf("|") != -1) s += DoArgs(args[j], "|", ins, outs); else s += args[j]; } } return s; } // Apply a definition to remaining rules function DoDefine(def, xx, i) { var sp = def.indexOf(" "); var eq = def.indexOf("=", sp + 1); if (sp == -1 || eq == -1) return; sp++; var ins = def.substr(sp, eq - sp); var outs = def.substr(eq + 1); t += "<br/>Replacing " + ins + " with " + outs; for (; i < xx.length; i++) if (xx[i].charAt(0) != '*') xx[i] = DoArgs(xx[i], " ", ins, outs); } // Generate a sentence. Uses globals o[], w[] function GenerateS(psr, xx) { // Always start with a bare S o = new Array(); o.push("S"); w = null; // Apply phrase structure rules till we have a sentence for (var r = 0; r < psr.length; r++) { if (Apply(psr[r])) { if (showpsr) t += "<br/>" + Stringize(o, w); o = oo; } } w = Lexicalize(o); t += "<p><b>Deep structure:</b> " + Stringize(o, w); // Apply transformations for (var x = 0; x < xx.length; x++) { var xform = xx[x]; if (xform.charAt(0) == '*') t += "<br/>" + xform; if (xform.substr(0,6) == "define") { DoDefine(xform, xx, x + 1); } else Xform(xform); } t += "<p><b>Surface structure:</b> " + Stringize(o, w); } // Embed one S (o/w) within another (oo/ww) at location tgt // Repalce o/w with the combined sentence function Embed(oo, ww, tgt) { var o2 = new Array(); var w2 = new Array(); for (var i = 0; i < tgt; i++) { o2.push(oo[i]); w2.push(ww[i]); } for (var i = 0; i < o.length; i++) { o2.push(o[i]); w2.push(w[i]); } for (var i = tgt + 1; i < oo.length; i++) { o2.push(oo[i]); w2.push(ww[i]); } o = o2; w = w2; } // User hit the Generate button function process() { //Read parameters theform = document.theform; var text = document.getElementById("mytext"); // Read the lexicon var s = theform.lex.value; lex = s.split("\n"); // Read the phrase structure rules s = theform.psr.value; psr = s.split("\n"); // Read the transformations s = theform.xform.value; xx = s.split("\n"); t = ""; embedding = false; inf = false; // Generate a sentence GenerateS(psr, xx); // Embedded sentences emb = FindFirst("S|Sinf"); if (emb != -1) { var oo = o; var ww = w; t += "<p><b>Creating embedded sentence</b>"; embedding = true; if (oo[emb] == "Sinf") { inf = true; oo[emb] = "S"; } GenerateS(psr, xx); Embed(oo, ww, emb); t += "<p><b>Embedded:</b> " + Stringize(o, w); } // Link to drawing program var url = "http://ironcreek.net/phpsyntaxtree/?" + wordbrax(o, w); t += '<p><a href="' + url + '" target="_blank"><b>Draw tree</b></a>'; // Final output t = "<b>Final:</b> " + Pronounce(w) + t; // Output text.innerHTML = t; } function helpme() { window.open("gtghelp.html"); } </script> <table width="100%"> <tr><td bgcolor="#EEC25A"> <h2><br>  <a href="kit.html"><img src="liltree.gif" border=0 align="absmiddle" height="53" width="60"></a> gtg</h2></td></tr> </td></tr> </table> <i>This is a Javascript program which allows you to define a generative grammar. Press Generate to test, and Help for documentation. <blockquote><span style="color:red;"><b>Warning:</b> This program is <b>in progress</b> and apt to change at any time. </span></blockquote> <p>Also see: <A href="markov.html">Markov generator</a>; <a href="ggg.html">Generative Grammar Gadget</a>; <a href="mg.html">Minimalism gadget</a>. <p>—Mark Rosenfelder, 2018</i> <hr> <table width="100%"> <tr> <td> Lexicon: <br><textarea id="lex" name="lex" rows="15" cols="30"> Q:Comp ø:Comp the:D this:D::p these that:D::p those cat:N::p cats dog:N::p dogs mouse:N::p mice man:N:m:p men woman:N:f:p women frog:N::p frogs monster:N::p monsters boy:N:m:p boys girl:N:f:p girls past:T: pres:T: not:Neg: must:Modal: can:Modal: have:Perf::pres has past had pastp had ing having en had be:Prog::pres is presp are past was pastp were ing being en been do:V:NP:pres does past did pastp did ing doing en done sit:V:PP:pres sits past sat pastp sat ing sitting en sat sleep:V:pres sleeps past slept pastp slept ing sleeping en slept hate:V:NP:pres hates past hated pastp hated ing hating en hated love:V:NP:pres loves past loved pastp loved ing loving en loved chase:V:NP:pres chases past chased pastp chased ing chasing en chased give:V:NP NP:pres gives past gave pastp gave ing giving en given think:V:S:pres thinks past thought pastp thought ing thinking en thought know:V:S:pres knows past knew pastp knew ing knowing en known want:V:Sinf:pres wants past wanted pastp wanted ing wanting en wanted put:V:NP PP:pres puts past put pastp put ing putting en put care:V::pres cares past cared pastp cared ing caring en cared by:P near:P to:P big:A little:A nice:A nasty:A beautiful:A ugly:A happy:A sad:A who:Pron:: he:Pron:m:acc him p they pacc them she:Pron:f:acc her p they pacc them everyone:Pron:: someone:Pron:: </textarea> </td> <td width="20px"> </td> <td> Phrase structure: <br><textarea id="psr" name="psr" rows="15" cols="30"> S=Comp TP TP=NP T VP VP=Modal VP=? VP=Perf VP=? VP=Prog VP=? VP=V (NP)|V NP PP|V S|V Sinf PP=P NP NP=D ?A N|Pron </textarea> </td> <td width="20px"> </td> <td> Transformations: <br><textarea id="xform" name="xform" rows="15" cols="55"> define Aux=Modal|Perf|Prog|Pass * suppress questions in subclauses sub lex ø if word Q paste * plural form of determiners start find next N copy num find previous D paste find next N loop * negative maybe if Aux lex not insert Neg maybe if no Aux find T lex not insert Neg * passive checkif V [NP [NP yes transform V [NP [NP into »be:Pass [VP 0 1 [PP »by 2 ] ] end subject copy object flip no transform V [NP into »be:Pass [VP 0 [PP »by 1 ] ] end subject copy object flip * wh-question main object ifword who cut prepend transform Comp into »Q main subject ifword who transform Comp into »ø * dative movement maybe transform [NP [NP into 1 [PP »to 0 ] * question formation if word Q transform [NP T into 1 0 if word Q find Aux cut find T paste * object forms of pronouns find V start find next Pron inflect acc loop if word inf find V find last Pron inflect acc * passive inflection if Pass find next V inflect en * progressive if Prog find next V|Pass inflect ing * perfect if Perf find next V|Pass|Prog inflect en * Infinitive S: suppress modal, add to if word inf find Modal cut if word inf find T lex to insert * Do support if no Aux lex do if Neg find T paste if no Aux lex do if no Neg if word Q find T paste * verbal inflection if no word inf if T cuttense numclip find V|Aux inflect clip </textarea> </td>  <td> <p><input type="button" value="Generate" onClick="process();"> <p><input type="button" value="Help me!" onClick="helpme();"> </td> </tr></table> <hr> <h3>Output</h3> <br><div id="mytext"> </div> </form> <hr> <center><A HREF="default.html"><img src="homeg.gif" border=0 alt="Home"></A></center> </body> </html>

CINXE.COM

gtg - Generative Tree Gadget